Medicine
MedHELM MedQA
MedHELM med_qa: per-(model, item) exact_match in {0,1} (1=correct) on 1000 US medical licensing exam (USMLE) multiple-choice clinical questions. 13 models from the MedHELM v4.0.0 release.
1,000items
13subjects
100%observed
Apache-2.0license
medicinedomain
textmodality
Response matrix
Fit to width. Hover for subject & item; click a cell for details.
Correct (1)Incorrect (0)Unobserved
Scale: 1 = correct · 0 = incorrect
Subjects
- 1openai/gpt-5-2025-08-070.966
- 2openai/gpt-5-mini-2025-08-070.957
- 3openai/o4-mini-2025-04-160.952
- 4google/gemini-2.5-pro-preview-05-060.942
- 5openai/o3-mini-2025-01-310.913
- 6openai/gpt-4o-2024-05-130.873
- 7deepseek-ai/deepseek-r10.869
- 8anthropic/claude-3-5-sonnet-202410220.855
- 9anthropic/claude-3-7-sonnet-202502190.854
- 10meta/llama-3.3-70b-instruct0.846
- 11google/gemini-2.0-flash-0010.831
- 12google/gemini-1.5-pro-0010.769
- 13openai/gpt-4o-mini-2024-07-180.761