MedScribe
As of Nov 3, 2026, MedScribe (v1) is maintained by Vals AI. Metric accuracy (%). Ceiling 100. Saturation unknown. Contamination risk low.
- Maintainer
- Vals AI
- Metric
- accuracy (%)
- Ceiling
- 100
- Saturation
- unknown
- Contamination
- low
- Cadence
- Declared cadence: as Vals updates (checked daily). Catalog snapshots: 1, latest Oct 30, 2026.
- Source
- Vals AI · this benchmark on Vals AI
- Grading
- AI judge (LLM-graded rubric) — SOAP notes from clinical conversations are scored by an LLM judge on expert rubrics. Dataset built with Protege.
Source: Vals AI. A proprietary private test set run by Vals; scores are republished with permission. Rows are dated the day isitsota first saw them, and only the benchmark's standard harness counts.
Categories
Where it counts
- Healthcare — secondary, weight 0.5
Top 5
Current top scores
- Claude Opus 5.5 · 91.43 · Oct 30, 2026
- Claude Fable 5.1 · 91.294 · Oct 30, 2026
- Claude Sonnet 5.5 · 91.101 · Oct 30, 2026
- Claude Opus 5 · 90.985 · Oct 30, 2026
- Muse Spark 1.2 · 90.062 · Oct 30, 2026
| Model | Score | Date | Kind |
|---|---|---|---|
| Claude Opus 5.5 | 91.43 | 2026-10-30 | snapshot |
| Claude Fable 5.1 | 91.294 | 2026-10-30 | snapshot |
| Claude Sonnet 5.5 | 91.101 | 2026-10-30 | snapshot |
| Claude Opus 5 | 90.985 | 2026-10-30 | snapshot |
| Muse Spark 1.2 | 90.062 | 2026-10-30 | snapshot |
| Claude Haiku 5.5 | 89.542 | 2026-10-30 | snapshot |
| Grok 4.7 | 89.377 | 2026-10-30 | snapshot |
| Muse Wonder Full | 89.335 | 2026-10-30 | snapshot |
| GLM 5.3 Flash | 88.936 | 2026-10-30 | snapshot |
| Muse Spark 1.1 | 88.888 | 2026-10-30 | snapshot |
| GLM 5.3 | 88.81 | 2026-10-30 | snapshot |
| Claude Fable 5 | 88.522 | 2026-10-30 | snapshot |
| Mimo V2.6 Pro | 88.307 | 2026-10-30 | snapshot |
| GPT-5.1 | 88.09 | 2026-10-30 | snapshot |
| Kimi K3 | 88.048 | 2026-10-30 | snapshot |
| GPT-6 Astra | 87.908 | 2026-10-30 | snapshot |
| Gemini 4 Argon | 87.427 | 2026-10-30 | snapshot |
| Minimax M3 | 87.252 | 2026-10-30 | snapshot |
| Grok 4.5 | 86.884 | 2026-10-30 | snapshot |
| GPT-5.5 | 86.868 | 2026-10-30 | snapshot |
| Claude Opus 4.6 | 86.738 | 2026-10-30 | snapshot |
| Grok 4.6 | 86.534 | 2026-10-30 | snapshot |
| GPT-6.1 Sol | 86.453 | 2026-10-30 | snapshot |
| Claude Opus 4.6 Thinking | 86.13 | 2026-10-30 | snapshot |
| Muse Spark | 85.902 | 2026-10-30 | snapshot |
| Claude Opus 4.8 | 85.755 | 2026-10-30 | snapshot |
| DeepSeek V4.1 Flash | 85.5 | 2026-10-30 | snapshot |
| Inkling | 85.405 | 2026-10-30 | snapshot |
| Claude Opus 4.5 20251101 Thinking | 85.321 | 2026-10-30 | snapshot |
| Mimo V2.6 Flash | 85.275 | 2026-10-30 | snapshot |