Medmarks - SuperGPQA Medicine Hard: leaderboard
The "SuperGPQA Medicine Hard" 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: Around February 2027. 89 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Pro (Preview) | 60.68 | #66 |
| 2 | Gemma 4 31B (IT) | 60.22 | #176 |
| 3 | Qwen 3.6 27B (Thinking) | 57.45 | #170 (Qwen 3.6 27B) |
| 4 | Qwen 3.6 35B A3B (Thinking) | 55.45 | #226 (Qwen 3.6 35B A3B) |
| 5 | Qwen 3.5 27B (Thinking) | 54.69 | #187 (Qwen 3.5 27B) |
| 6 | GPT-5.1 (Medium) | 52.38 | #135 (GPT-5.1) |
| 7 | Qwen 3.5 35B A3B (Thinking) | 49.77 | #256 (Qwen 3.5 35B A3B) |
| 8 | Qwen 3 Next 80B A3B (Thinking) | 49.46 | #320 (Qwen 3 Next 80B A3B) |
| 9 | GLM-4.7 FP8 | 48.39 | #209 |
| 10 | GPT-5.2 (Medium) | 48.23 | #110 (GPT-5.2) |
| 11 | Claude Sonnet 4.5 | 47.93 | #142 |
| 12 | MiniMax-M2.1 | 47.77 | #289 |
| 13 | Grok 4 | 46.85 | #181 |
| 14 | GPT-OSS-120B (High) | 46.7 | #341 (GPT-OSS-120B) |
| 15 | Qwen 3 30B A3B (Thinking) | 46.39 | #493 (Qwen 3 30B 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-supergpqa-medicine-hard · How It Works · Data refreshed daily, snapshot 2026-10-20.