Medmarks: leaderboard
Medical model evaluations across verifiable and open-ended tasks. This card reports the verifiable weighted win rate, based on pairwise comparison points; open-ended results are a separate board.
Metric: Weighted Mean Win Rate (%). Source: medmarks.ai. Saturation forecast: Around February 2027. 89 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Pro (Preview) | 64.52 | #66 |
| 2 | GPT-5.1 (Medium) | 62.08 | #135 (GPT-5.1) |
| 3 | Grok 4 | 61.59 | #181 |
| 4 | Claude Sonnet 4.5 | 60.71 | #142 |
| 5 | GPT-5.2 (Medium) | 60.5 | #110 (GPT-5.2) |
| 6 | GLM-4.7 FP8 | 60.1 | #209 |
| 7 | Qwen 3.5 27B (Thinking) | 59.33 | #187 (Qwen 3.5 27B) |
| 8 | Qwen 3.6 35B A3B (Thinking) | 59.32 | #226 (Qwen 3.6 35B A3B) |
| 9 | Gemma 4 31B (IT) | 59.05 | #176 |
| 10 | Qwen 3.5 35B A3B (Thinking) | 58.56 | #256 (Qwen 3.5 35B A3B) |
| 11 | Qwen 3.6 27B (Thinking) | 58.5 | #170 (Qwen 3.6 27B) |
| 12 | Qwen 3 235B A22B (Thinking) | 58.42 | #313 (Qwen 3 235B A22B) |
| 13 | Muse Glimmer 30B (xHigh) | 57.59 | #178 (Muse Glimmer 30B) |
| 14 | Muse Glimmer 30B (High) | 57.43 | #178 (Muse Glimmer 30B) |
| 15 | Qwen 3 Next 80B A3B (Thinking) | 56.95 | #320 (Qwen 3 Next 80B A3B) |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=medmarks · How It Works · Data refreshed daily, snapshot 2026-10-20.