Medmarks - SuperGPQA Medicine Hard: leaderboard
Metric: Score (%). Source: medmarks.ai. 71 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3 Pro (Preview) | 60.68 |
| 2 | GPT-5.1 (Medium) | 52.38 |
| 3 | Qwen 3 Next 80B A3B (Thinking) | 49.46 |
| 4 | GLM-4.7 FP8 | 48.39 |
| 5 | GPT-5.2 (Medium) | 48.23 |
| 6 | Claude Sonnet 4.5 | 47.93 |
| 7 | MiniMax-M2.1 | 47.77 |
| 8 | Grok 4 | 46.85 |
| 9 | GPT-OSS-120B (High) | 46.7 |
| 10 | Qwen 3 Next 80B A3B Instruct | 46.24 |
| 11 | Qwen 3 235B A22B (Thinking) | 45.93 |
| 12 | GPT-OSS-120B (Medium) | 45.01 |
| 13 | Qwen 3 VL 30B A3B (Thinking) | 43.63 |
| 14 | INTELLECT-3 | 43.32 |
| 15 | GPT-OSS-120B (Low) | 41.94 |
Interactive version: theaggregate.ai/benchmark?slug=medmarks-supergpqa-medicine-hard · How It Works · Data refreshed daily, snapshot 2026-09-05.