PhysicianBench: leaderboard
Long-horizon physician workflow benchmark grounded in clinical records, measuring checkpoint and end-to-end task success.
Metric: Pass@1 (imported). Source: healthrex.github.io. Source: benchmarklist.com. Saturation forecast: Around July 2027. 12 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.5 | 46.3 | #31 |
| 2 | Claude Opus 4.6 | 31.7 | #63 |
| 3 | Claude Opus 4.7 | 29.3 | #49 |
| 4 | GPT-5.4 | 27.7 | #79 |
| 5 | Claude Sonnet 4.6 | 23 | #90 |
| 6 | DeepSeek V4 Pro | 18.7 | #102 |
| 7 | Kimi K2.6 | 17 | #104 |
| 8 | MiMo-V2.5-Pro | 16.7 | #158 |
| 9 | Qwen 3.6 Plus | 13.7 | #120 |
| 10 | MiniMax-M2.7 | 8.7 | #258 |
| 11 | Gemini 3.1 Pro (Preview) | 6 | #58 |
| 12 | Grok 4.20 | 5.3 | #175 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=physicianbench · How It Works · Data refreshed daily, snapshot 2026-10-20.