Medmarks - CareQA EN: leaderboard
The "CareQA EN" 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) | 96.8 | #66 |
| 2 | Grok 4 | 95.07 | #181 |
| 3 | GPT-5.1 (Medium) | 94.91 | #135 (GPT-5.1) |
| 4 | GPT-5.2 (Medium) | 94.69 | #110 (GPT-5.2) |
| 5 | GLM-4.7 FP8 | 94.35 | #209 |
| 6 | Claude Sonnet 4.5 | 93.54 | #142 |
| 7 | Qwen 3.5 27B (Thinking) | 93.44 | #187 (Qwen 3.5 27B) |
| 8 | Qwen 3 235B A22B (Thinking) | 93.13 | #313 (Qwen 3 235B A22B) |
| 9 | Qwen 3.5 35B A3B (Thinking) | 92.77 | #256 (Qwen 3.5 35B A3B) |
| 10 | Qwen 3.6 27B (Thinking) | 92.76 | #170 (Qwen 3.6 27B) |
| 11 | Qwen 3.6 35B A3B (Thinking) | 92.4 | #226 (Qwen 3.6 35B A3B) |
| 12 | Gemma 4 31B (IT) | 92.29 | #176 |
| 13 | Qwen 3 Next 80B A3B (Thinking) | 92.14 | #320 (Qwen 3 Next 80B A3B) |
| 14 | MiniMax-M2.1 | 91.84 | #289 |
| 15 | Qwen 3 Next 80B A3B Instruct | 91.61 | #368 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=medmarks-careqa-en · How It Works · Data refreshed daily, snapshot 2026-10-20.