Skip to content

HealthBench Professional

An open benchmark for clinician-facing model responses across care consult, writing and documentation, and medical research tasks.

16 models

Higher is better.

  1. 1Claude Sonnet 5.5Anthropic69.2
  2. 2Claude Opus 5.5Anthropic65.6
  3. 3Ling 3.1 FlashInclusionAI65.3
  4. 4GPT-6 AstraOpenAI64.7
  5. 5GPT-6.1 SolOpenAI64.2
  6. 6Muse Wonder FullMeta62.3
  7. 7GPT-6 SolOpenAI60.8
  8. 8GPT-6 LunaOpenAI60.8
  9. 9GPT-5.6 SolOpenAI60.5
  10. 10Muse WonderMeta60.5
  11. 11Claude Opus 5Anthropic59.8
  12. 12Muse Spark 1.1Meta59.3
  13. 13GPT-5.6 TerraOpenAI57.7
  14. 14Grok 4.7xAI56.7
  15. 15GPT-5.6 LunaOpenAI55.7
  16. 16GPT-5.4OpenAI48.1
Higher is better. Bar colour: blue proprietary, orange open weights, green other.

About this benchmark

Metric
Rubric-graded open-ended responses
Tasks
Clinician chat tasks
Authors
Rebecca Soskin Hicks, Mikhail Trofimov, Dominick Lim, Rahul K. Arora, Foivos Tsimpourlas, Preston Bowman, Michael Sharman, Chi Tong, Kavin Karthik, Arnav Dugar, Akshay Jagadeesh, Khaled Saab, Johannes Heidecke, Ashley Alexander, Nate Gross, Karan Singhal (2026)

Scores are as reported by model developers or evaluators and may use different settings (for example reasoning effort or tools). Compare with care.