Vals AI MedScribe: leaderboard
Generating structured SOAP notes from synthetic doctor-patient conversation transcripts, scored against rubrics created by expert medical scribes and physicians.
Metric: Accuracy (%). Source: www.vals.ai. Source: benchmarklist.com. Saturation forecast: Around January 2027. 109 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 5.5 (Max) | 91.43 | #3 (Claude Opus 5.5) |
| 2 | Claude Fable 5.1 (Max) | 91.29 | #9 (Claude Fable 5.1) |
| 3 | Claude Sonnet 5.5 (Max) | 91.1 | #11 (Claude Sonnet 5.5) |
| 4 | Claude Opus 5 (Max) | 90.98 | #12 (Claude Opus 5) |
| 5 | Muse Spark 1.2 (xHigh) | 90.06 | #62 (Muse Spark 1.2) |
| 6 | Claude Haiku 5.5 (Max) | 89.54 | #40 (Claude Haiku 5.5) |
| 7 | Grok 4.7 (xHigh) | 89.38 | #25 (Grok 4.7) |
| 8 | Muse Wonder Full (Max) | 89.33 | #10 (Muse Wonder Full) |
| 9 | GLM-5.3 Flash (Max) | 88.94 | #65 (GLM-5.3 Flash) |
| 10 | Muse Spark 1.1 (xHigh) | 88.89 | #52 (Muse Spark 1.1) |
| 11 | GLM-5.3 (Max) | 88.81 | #39 (GLM-5.3) |
| 12 | Claude Fable 5 (Max) | 88.52 | #16 (Claude Fable 5) |
| 13 | MiMo-V2.6-Pro | 88.31 | #42 |
| 14 | GPT-5.1 (2025-11-13) (High) | 88.09 | #101 (GPT-5.1 (2025-11-13)) |
| 15 | Kimi K3 (Max) | 88.05 | #30 (Kimi K3) |
Interactive version: theaggregate.ai/benchmark?slug=vals-ai-medscribe · How It Works · Data refreshed daily, snapshot 2026-10-20.