Which AI model scores highest on HealthBench Professional?

GPT-6 Astra by OpenAI holds the best HealthBench Professional (OpenAI grading, raw) result among tracked models, at 68.2% (released Sep 3 2026). Higher scores are better on this benchmark.

What are the top 3 models on HealthBench Professional?

1. GPT-6 Astra (OpenAI) — 68.2%; 2. GPT-6 Luna (OpenAI) — 61.2%; 3. GPT-6 Sol (OpenAI) — 59.5%.

How many models have a published HealthBench Professional score?

3 tracked models have a published HealthBench Professional (OpenAI grading, raw) score. Scores are collected from published lab reports and benchmark sources, with recorded sources credited alongside the results. We do not re-run the evaluations.

Health

HealthBench ProfessionalOpenAI grading, raw

unadjusted physician-rubric score using the official OpenAI grading procedure

HealthBench Professional raw OpenAI grading, raw OpenAI grading, length-adjusted length-adjusted

Rankings

Higher is better

HealthBench Professional — frequently asked questions

: unadjusted physician-rubric score using the official OpenAI grading procedure
: GPT-6 Astra by OpenAI holds the best HealthBench Professional (OpenAI grading, raw) result among tracked models, at 68.2% (released Sep 3 2026). Higher scores are better on this benchmark.
: 1. GPT-6 Astra (OpenAI) — 68.2%; 2. GPT-6 Luna (OpenAI) — 61.2%; 3. GPT-6 Sol (OpenAI) — 59.5%.
: 3 tracked models have a published HealthBench Professional (OpenAI grading, raw) score. Scores are collected from published lab reports and benchmark sources, with recorded sources credited alongside the results. We do not re-run the evaluations.