Medmarks - MedQA: leaderboard
The "MedQA" subtask of Medmarks. Living medical-model benchmark spanning verifiable and open-ended clinical evaluation families with model/system comparisons.
Metric: Score (%). Source: medmarks.ai. Saturation forecast: Estimated already saturated. 89 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Pro (Preview) | 95.39 | #66 |
| 2 | GLM-4.7 FP8 | 94.79 | #209 |
| 3 | Qwen 3.5 27B (Thinking) | 94.61 | #187 (Qwen 3.5 27B) |
| 4 | Qwen 3.5 35B A3B (Thinking) | 93.72 | #256 (Qwen 3.5 35B A3B) |
| 5 | Grok 4 | 93.48 | #181 |
| 6 | GPT-OSS-120B (High) | 93.27 | #341 (GPT-OSS-120B) |
| 7 | Qwen 3 235B A22B (Thinking) | 92.96 | #313 (Qwen 3 235B A22B) |
| 8 | Gemma 4 31B (IT) | 92.93 | #176 |
| 9 | Claude Sonnet 4.5 | 92.83 | #142 |
| 10 | Qwen 3.6 35B A3B (Thinking) | 91.86 | #226 (Qwen 3.6 35B A3B) |
| 11 | GPT-OSS-120B (Medium) | 91.54 | #341 (GPT-OSS-120B) |
| 12 | Qwen 3 Next 80B A3B (Thinking) | 91.12 | #320 (Qwen 3 Next 80B A3B) |
| 13 | MiniMax-M2.1 | 91.1 | #289 |
| 14 | MiniMax-M2 | 90.7 | #317 |
| 15 | INTELLECT-3 | 90.52 | #427 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=medmarks-medqa · How It Works · Data refreshed daily, snapshot 2026-10-20.