HealthBench
Health conversations scored against physician-written rubrics for accuracy, safety, communication, and context-seeking.
Failure modes HealthBench catches
Unsafe advice
Missing context
Poor escalation
Starter command
A starting point, not a guarantee. Commands change between releases, so confirm against the official README.
pip install inspect-ai inspect-evals
inspect eval inspect_evals/healthbench --model openai/gpt-4.1 --limit 20Sources
- 1. HealthBench official page or repository · OpenAI · Repository · Sources checked
- 2. HealthBench paper · arXiv · Paper · Sources checked