LLMEval-Med - Medical Text Generation: leaderboard
LLMEval-Med medical writing tasks (rewriting, generation and summarization of clinical documents), mostly in Chinese. Physicians rate instruction following, correctness, effectiveness, readability and safety, and a response is usable only above the combined threshold with no safety failure; usability rate in percent, higher is better. Zhang et al. (Fudan University), Findings of EMNLP 2025.
Metric: Usability rate: share of responses physicians rate usable on five dimensions with a safety veto (%). Source: arxiv.org. Saturation forecast: Estimated already saturated. 13 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o | 64.58 | #741 |
| 2 | DeepSeek V3 | 49.48 | #705 |
| 3 | O1 Mini | 49.48 | #761 |
| 4 | O1 Preview | 49.48 | #483 |
| 5 | DeepSeek R1 | 44.33 | #467 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=llmeval-med-medical-text-generation · How It Works · Data refreshed daily, snapshot 2026-11-12.