{"schema_version":1,"total":1,"offset":0,"limit":100,"observations":[{"id":"public:3a000bc53a2323936aaf9aa1","benchmark_id":"surge-gdp-pdf::100-tasks","subject":{"source_id":"Mistral Large 3","name":"Mistral Large 3","model_id":"mistral-large-3::default","variant":null,"harness":null},"value":2,"unit":"percent","basis":"measured","source":{"url":"https://surgehq.ai/benchmarks/gdp-pdf","retrieved_at":"2026-09-22T08:23:55.126460+00:00","published_at":null,"file":"data/raw/benchmarks/daily-evidence/2026-09-22-surge/d44bb04fd9aa6e63f5c8.gz","sha256":"d44bb04fd9aa6e63f5c838020028d9100164cf3acebe5327feb9b711c5cda972","locator":"surge_benchmark_board; source row 38; Mistral Large 3; field value"},"protocol":"GDP.pdf (Surge AI): 100 held-out tasks, each a real professional prompt with its PDF documents, no tools; a response passes only when every rubric criterion is satisfied (all_pass), judged by Gemini 3.5 Flash; leaderboard configuration five runs per task; value = the board's all-pass rate (percent) for the row's model and stated setting; Surge runs the models itself; source row: {\"brand\":\"Mistral\",\"stated_setting\":\"not stated\",\"board\":\"GDP.pdf\"}","comparison_key":null,"supporting_sources":[{"url":"https://raw.githubusercontent.com/surge-ai/gdp-pdf/main/README.md","file":"data/raw/benchmarks/daily-evidence/2026-09-22-surge/fda47ebf36bf4299b5d6.gz","sha256":"fda47ebf36bf4299b5d69288978de80e06fa6a001b8bc75550eb1f38d8e9e148","retrieved_at":"2026-09-22T08:24:00.322672+00:00","published_at":null,"locator":"method_source"}],"join_note":"Reviewed identity map 2026-09-22: label names mistral-large-3 without a setting; the catalog has exactly one configuration, the default (mistral-large-3::default)"}],"divergences":[],"cell":null,"coverage":{"model":null,"benchmark":{"total_models":924,"available":36,"unknown":888,"not_tested":0,"not_published":0,"source_unreachable":0,"contested":0,"measured":36,"self_reported":0,"observations":53,"unmatched_observations":17}},"collection":{"benchmark_id":"surge-gdp-pdf::100-tasks","status":"collected","source_url":"https://surgehq.ai/benchmarks/gdp-pdf","reason":"41 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."}