EHRBench - PROMOTE: leaderboard
EHRBench (2026) tests LLMs on clinical decision questions generated from real patient records and verified against biomedical knowledge bases, posed as multiple-choice questions with 4 to 6 options. This board is accuracy in percent on the questions built from PROMOTE records; higher is better.
Metric: Accuracy (%) on the questions built from PROMOTE records, EHR-grounded multiple-choice clinical decision questions (4, 5 and 6 options) built from MIMIC-III, MIMIC-IV and PROMOTE encounter trajectories with knowledge-base verification, JSON-constrained answers, deterministic decoding; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 31 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.2 Instant | 70.7 | #309 |
| 2 | GPT-5 Chat | 68.26 | #452 |
| 3 | GPT-4.1 | 67.87 | #506 |
| 4 | GLM-4 32B (0414) | 65.85 | #1525 |
| 5 | Qwen 3 32B | 65.55 | #983 |
| 6 | Llama 3.3 70B Instruct | 65.35 | #1394 |
| 7 | GPT-4.1 Mini | 64.45 | #799 |
| 8 | GPT-5 Mini | 63.88 | #373 |
| 9 | Mistral Small 3 | 63.38 | #1809 |
| 10 | Qwen 2.5 32B | 63.05 | #1347 |
| 11 | m1-32B-1k | 61.46 | #1247 |
| 12 | EHRBench llama3-70B (checkpoint unspecified) | 61.2 | |
| 13 | doctor-r1-8B | 59.01 | #1634 |
| 14 | Qwen 3 8B | 58.74 | #1890 |
| 15 | Qwen 3 4B | 58.39 | #2379 |
No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=ehrbench-promote · How It Works · Data refreshed daily, snapshot 2026-11-22.