Medicine
MedHELM EHRSQL
MedHELM ehr_sql: per-(model, item) ehr_sql_execution_accuracy in {0,1} (1=correct) on 1000 translating a natural-language clinical question into SQL over an EHR database. 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-4o-2024-05-130.32
- 2openai/o4-mini-2025-04-160.29
- 3openai/o3-mini-2025-01-310.272
- 4openai/gpt-5-2025-08-070.233
- 5google/gemini-2.5-pro-preview-05-060.213
- 6openai/gpt-5-mini-2025-08-070.193
- 7google/gemini-2.0-flash-0010.165
- 8google/gemini-1.5-pro-0010.152
- 9anthropic/claude-3-5-sonnet-202410220.14
- 10openai/gpt-4o-mini-2024-07-180.112
- 11anthropic/claude-3-7-sonnet-202502190.081
- 12deepseek-ai/deepseek-r10.076
- 13meta/llama-3.3-70b-instruct0.074