Knowledge and exam benchmarks

1 tracked · snapshot reviewed September 28, 2026

These evaluations test medical knowledge and reasoning through questions with reference answers. Some include clinical images and specialist-level material. Performance on a question set does not directly measure a model's ability to deliver care in practice.

MedXpertQA (MM)

Tsinghua University · 2,000 multimodal questions

Expert-level multiple-choice questions over clinical images across 17 specialties.

81.5
GPT-5.6 Sol partial review
81.3%
Gemini 3.1 Pro
80.4%
A
Qwen3.8 Max partial review
80.0
Claude Fable 5 partial review
78.4%
Muse Spark
77.1%
GPT-5.4
76.0%
Gemini 3 Pro
73.3
GPT-5.2
71.7
Claude Opus 4.8 partial review
71.0%
A
Qwen3.7 Plus partial review
70.0
A
Qwen3.5 397B A17B
68.7
A
Qwen3.6 Plus partial review

Showing top 12 of 22 indexed results. View all results.

Which knowledge and exam benchmarks have results in this index?

1 as of September 28, 2026: MedXpertQA (MM) (GPT-5.6 Sol: highest indexed score 81.5).

The other categories sit on the index: rubric-graded benchmarks, agentic and workflow benchmarks, documentation and coding benchmarks, safety benchmarks, composite indices.