Medmarks - PubHealthBench Reviewed — leaderboard
Metric: Score (%). Source: medmarks.ai. 71 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3 Pro (Preview) | 94.82 |
| 2 | GPT-5.2 (Medium) | 94.08 |
| 3 | GPT-5.1 (Medium) | 93.6 |
| 4 | Grok 4 | 91.71 |
| 5 | Claude Sonnet 4.5 | 91.49 |
| 6 | GLM-4.7 FP8 | 90.7 |
| 7 | MiniMax-M2.1 | 90.61 |
| 8 | GPT-OSS-120B (High) | 88.95 |
| 9 | Llama 3.3 70B Instruct | 88.25 |
| 10 | GPT-OSS-120B (Medium) | 88.25 |
| 11 | GPT-OSS-120B (Low) | 87.89 |
| 12 | Qwen 3 Next 80B A3B Instruct | 87.68 |
| 13 | Qwen 3 235B A22B (Thinking) | 87.41 |
| 14 | Ling-flash-2.0 | 86.49 |
| 15 | Hermes 4 70B | 86.32 |
Interactive version: theaggregate.ai/benchmark?slug=medmarks-pubhealthbench-reviewed · How the rankings work · Data refreshed daily, snapshot 2026-07-22.