HealthBench

OpenAI's HealthBench: 5,000 health conversations graded against physician-written rubrics, from emergency referrals to global health.

Score distribution

Include deprecated models

Rank	Model	Score
1	Claude Sonnet 5.5 (Max, Default Fallback) Anthropic	62.1%
2	Claude Fable 5.2 (Low, Default Fallback) Anthropic	60.4%
3	Claude Fable 5.2 (Medium, Default Fallback) Anthropic	60.2%
4	Claude Fable 5.2 (High, Default Fallback) Anthropic	59.8%
5	Claude Fable 5.2 (Xhigh, Default Fallback) Anthropic	59.7%
6	Claude Fable 5.2 (Max, Default Fallback) Anthropic	59.6%
7	GPT-6.1 Sol (Max) OpenAI	58.5%
8	Claude Haiku 5.5 (Max) Anthropic	58.5%
9	GPT-6 Astra (Max) OpenAI	58.2%
10	Claude Opus 5.5 (Max, Default Fallback) Anthropic	57.0%
11	Grok 4.7 (Xhigh) SpaceXAI	56.0%
12	GPT-6 Luna (Max) OpenAI	54.6%
13	GLM-5.3 (Max) Z AI	54.5%
14	Gemini 4 Argon (High) Google	49.1%
15	Gemini 3.8 Flash (High) Google	48.4%