{"source":"CDSBench","benchmarkVersion":"v0.1","url":"https://cdsbench.com/leaderboard","generatedAt":"2026-09-16T11:28:51.739Z","license":"Public benchmark results may be reused with attribution to CDSBench and a link to the source URL. Source documents are not included.","note":"Frontier models are ranked. Mastermind is CDSBench's own system built on them and is included as a reference row, never ranked.","models":[{"model":"Mastermind","version":"Mastermind V0 (raw baseline)","vendor":"CDSBench","ranked":false,"overall":91.2223,"accuracy":0.9054,"evidence_support":0.9226,"multi_document_reasoning":0.7273,"uncertainty_handling":0.9392,"hallucinated_evidence_rate":0.0608,"false_high_confidence_rate":0,"questions":296,"change_since_previous_run":0,"run_finished_at":"2026-09-15T20:36:13.712Z","byCategory":{"extraction":{"accuracy":1,"n":56,"correct":56},"maturity":{"accuracy":1,"n":26,"correct":26},"currency":{"accuracy":1,"n":28,"correct":28},"seniority":{"accuracy":1,"n":30,"correct":30},"transferability":{"accuracy":0.5666666666666667,"n":30,"correct":17},"contradiction":{"accuracy":1,"n":2,"correct":2},"deliverability":{"accuracy":0.8333333333333334,"n":30,"correct":25},"uncertainty":{"accuracy":0.6875,"n":32,"correct":22},"restructuring":{"accuracy":1,"n":2,"correct":2},"entity_match":{"accuracy":1,"n":30,"correct":30},"evidence":{"accuracy":1,"n":30,"correct":30}}},{"model":"Claude","version":"Claude Sonnet 5","vendor":"Anthropic","ranked":true,"overall":91.2223,"accuracy":0.9054,"evidence_support":0.9226,"multi_document_reasoning":0.7273,"uncertainty_handling":0.9392,"hallucinated_evidence_rate":0.0608,"false_high_confidence_rate":0,"questions":296,"change_since_previous_run":0,"run_finished_at":"2026-09-15T20:36:08.094Z","byCategory":{"extraction":{"accuracy":1,"n":56,"correct":56},"maturity":{"accuracy":1,"n":26,"correct":26},"currency":{"accuracy":1,"n":28,"correct":28},"seniority":{"accuracy":1,"n":30,"correct":30},"transferability":{"accuracy":0.5666666666666667,"n":30,"correct":17},"contradiction":{"accuracy":1,"n":2,"correct":2},"deliverability":{"accuracy":0.8333333333333334,"n":30,"correct":25},"uncertainty":{"accuracy":0.6875,"n":32,"correct":22},"restructuring":{"accuracy":1,"n":2,"correct":2},"entity_match":{"accuracy":1,"n":30,"correct":30},"evidence":{"accuracy":1,"n":30,"correct":30}}},{"model":"GPT","version":"GPT-5","vendor":"OpenAI","ranked":true,"overall":61.4925,"accuracy":0.8108,"evidence_support":0.212,"multi_document_reasoning":0.7879,"uncertainty_handling":0.3333,"hallucinated_evidence_rate":0.1486,"false_high_confidence_rate":0.1892,"questions":296,"change_since_previous_run":0,"run_finished_at":"2026-09-15T20:36:05.308Z","byCategory":{"extraction":{"accuracy":0.9642857142857143,"n":56,"correct":54},"maturity":{"accuracy":1,"n":26,"correct":26},"currency":{"accuracy":1,"n":28,"correct":28},"seniority":{"accuracy":1,"n":30,"correct":30},"transferability":{"accuracy":0.9333333333333333,"n":30,"correct":28},"contradiction":{"accuracy":1,"n":2,"correct":2},"deliverability":{"accuracy":0.7333333333333333,"n":30,"correct":22},"uncertainty":{"accuracy":0.6875,"n":32,"correct":22},"restructuring":{"accuracy":0,"n":2,"correct":0},"entity_match":{"accuracy":0.9333333333333333,"n":30,"correct":28},"evidence":{"accuracy":0,"n":30,"correct":0}}},{"model":"Gemini","version":"Gemini 2.5 Pro","vendor":"Google","ranked":true,"overall":58.6194,"accuracy":0.777,"evidence_support":0.2119,"multi_document_reasoning":0.8788,"uncertainty_handling":0.3333,"hallucinated_evidence_rate":0.2128,"false_high_confidence_rate":0.223,"questions":296,"change_since_previous_run":0,"run_finished_at":"2026-09-15T20:36:10.931Z","byCategory":{"extraction":{"accuracy":0.6785714285714286,"n":56,"correct":38},"maturity":{"accuracy":1,"n":26,"correct":26},"currency":{"accuracy":1,"n":28,"correct":28},"seniority":{"accuracy":1,"n":30,"correct":30},"transferability":{"accuracy":1,"n":30,"correct":30},"contradiction":{"accuracy":1,"n":2,"correct":2},"deliverability":{"accuracy":0.7333333333333333,"n":30,"correct":22},"uncertainty":{"accuracy":0.6875,"n":32,"correct":22},"restructuring":{"accuracy":1,"n":2,"correct":2},"entity_match":{"accuracy":1,"n":30,"correct":30},"evidence":{"accuracy":0,"n":30,"correct":0}}}]}