MedScribe
Professional Work · 2026-02-19
Clinical documentation benchmark testing whether models can help doctors with administrative writing tasks such as notes, summaries, and structured medical documentation.
Top models (higher is better)
| Model | Score |
|---|---|
| Claude Opus 5 | 91.0 |
| Muse Spark 1.1 | 88.9 |
| Claude Fable 5 | 88.5 |
| GPT-5.1 | 88.1 |
| Kimi K3 | 88.0 |
| MiniMax M3 | 87.3 |
| Grok 4.5 | 86.9 |
| GPT-5.5 | 86.9 |
| Opus 4.6 | 86.7 |
| Muse Spark | 85.9 |
| Opus 4.8 | 85.8 |
| Inkling | 85.4 |
| Opus 4.5 | 85.3 |
| GPT-5.6 Sol | 85.2 |
| Haiku 4.5 | 85.2 |
| Sonnet 4.5 | 84.5 |
| GPT-5.6 Luna | 84.4 |
| GPT-5.2 | 84.4 |
| MiMo-V2.5-Pro | 83.7 |
| GPT-5 | 83.7 |
| GLM-5.2 | 83.5 |
| Gemini 2.5 Flash | 83.0 |
| Opus 4.7 | 83.0 |
| GPT-5.6 Terra | 82.9 |
| Grok 4 Fast | 81.6 |
| MiniMax M2.1 | 80.8 |
| GPT-5 Mini | 80.6 |
| MiniMax M2.7 | 79.9 |
| Gemini 3.6 Flash | 79.7 |
| Qwen3.7-Max | 79.4 |
| Grok 4.1 Fast | 78.7 |
| Gemini 2.5 Flash Preview (09-2025) | 78.5 |
| Grok 4 | 78.2 |
| Kimi K2.6 | 78.1 |
| GPT-5.4 | 77.5 |
| Qwen3-VL-Plus-2025-09-23 | 77.1 |
| GPT-5.4 Nano | 77.1 |
| Qwen3.6 Plus (2026-04-02) | 77.0 |
| o3 | 76.7 |
| Gemini 3.5 Flash | 76.6 |