ResearchBenchmarksMedical Question Answering on HealthBench ConsensusFor AI assistantsMarkdownContent indexMedical Question Answering on HealthBench ConsensusFollow90.9ScoreGPT-5.281.64484.04786.4588.853Jun 21, 2026Evaluation resultsMethodMethodLinksScoreGPT-5.22026.0690.9Reported2026.0688Ideal completions2026.0682