{"schema_version":1,"total":1,"offset":0,"limit":100,"observations":[{"id":"public:0ea3cf191b80cdaa6393c40d","benchmark_id":"frontiermath-tier-4::v2","subject":{"source_id":"deepseek-v4-pro_max","name":"deepseek-v4-pro_max","model_id":"deepseek-v4-pro::max","variant":null,"harness":null},"value":0.024390243902439025,"unit":"fraction","basis":"measured","source":{"url":"https://epoch.ai/data/benchmark_data.zip","retrieved_at":"2026-09-16T09:43:00Z","published_at":null,"file":"data/raw/benchmarks/daily-evidence/2026-09-16-epoch-hub/epoch-frontiermath_tier_4_v2.csv.gz","sha256":"0c5a60359ff520a571a00b1c9d0001893d950607b5f7085ca96add349c6e0ca9","locator":"csv; source row 32; deepseek-v4-pro_max; field Best score (across scorers)"},"protocol":"FrontierMath Tier 4 v2 (Epoch AI) as published in Epoch AI's Benchmarking Hub (frontiermath_tier_4_v2.csv); value = Best score (across scorers), a fraction; one row per Epoch model version; not a Composite input.; mean_score=0.024390243902439025; stderr=0.024390243902439025; Started at=2026-06-17T02:30:14.000Z; id=RAZ6VMfGtQSE8EvUZRRz4Q","comparison_key":null,"join_note":"Reviewed identity map 2026-09-16: label states model deepseek-v4-pro and effort max; exact catalog configuration deepseek-v4-pro::max"}],"divergences":[],"cell":null,"coverage":{"model":null,"benchmark":{"total_models":844,"available":33,"unknown":811,"not_tested":0,"not_published":0,"source_unreachable":0,"contested":0,"measured":33,"self_reported":0,"observations":62,"unmatched_observations":29}},"collection":{"benchmark_id":"frontiermath-tier-4::v2","status":"collected","source_url":"https://epoch.ai/data/benchmark_data.zip","reason":"62 source results parsed; configurations remain separate; unmatched model identities are retained."},"coverage_note":"Denominators are catalog model configurations × versioned registry entries; capability_available excludes Efficiency (cost) boards. Basis counts count covered cells, not runs; measured and self_reported may overlap. Unmatched source identities remain in observations and never inflate catalog coverage."}