HealthBench AIHealth AI evaluation research
HealthBenchAI model evaluations and benchmark guides, published by REAGEN, Inc. Read our results, compare published scores, inspect supporting evidence and understand differences in test settings.
Claude Fable 5.1: our Professional test report
Our completed 525-task HealthBench Professional test: scores, confidence interval, model settings and downloadable per-task evidence.
Read article and sources →GPT-6.1 Sol Max: our Professional test report
Our completed 525-task GPT-6.1 Sol Max test: 64.63 length-adjusted, 67.56 raw, confidence interval, protocol and downloadable per-task evidence.
Read article and sources →Reading HealthBench Professional results
Our guide to clinical conversation scores, length adjustment and why evaluation protocols matter.
Read article and sources →Nox Orion: our test results
We tested Nox Orion and obtained these scores: 69.53 on HealthBench Professional and 30.79 on HealthBench Hard, both length-adjusted.
Read article and sources →HealthBench and Hard: a guide to the distinction
How we interpret the 5,000-conversation HealthBench dataset and its 1,000-example difficult subset.
Read article and sources →PubMedQA: interpreting yes, no and maybe
Our guide to abstract-based biomedical question answering, test splits and comparable PubMedQA scores.
Read article and sources →A benchmark score is not clinical validation. Different tools, graders, reasoning settings and data splits can make results non-comparable. Read our methodology.