Medmarks - Med-HALT Reasoning NOTA: leaderboard
The "Med-HALT Reasoning NOTA" 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 January 2027. 89 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3 Next 80B A3B (Thinking) | 78.86 | #320 (Qwen 3 Next 80B A3B) |
| 2 | Grok 4 | 78.54 | #181 |
| 3 | Gemini 3 Pro (Preview) | 74.4 | #66 |
| 4 | Qwen 3 14B (Reasoning) | 73.78 | #537 (Qwen 3 14B) |
| 5 | OLMo 3.1 32B (Thinking) | 72.07 | #632 (OLMo 3.1 32B) |
| 6 | Muse Glimmer 30B (xHigh) | 71.63 | #178 (Muse Glimmer 30B) |
| 7 | Muse Glimmer 30B (High) | 70.77 | #178 (Muse Glimmer 30B) |
| 8 | GPT-5.1 (Medium) | 69.85 | #135 (GPT-5.1) |
| 9 | GPT-5.2 (Medium) | 69.75 | #110 (GPT-5.2) |
| 10 | Nemotron 3.5 Lightning | 69.19 | #467 |
| 11 | GLM-4.7 FP8 | 69 | #209 |
| 12 | Qwen 2.5 32B Instruct | 68.78 | #498 |
| 13 | Qwen 3.5 27B (Thinking) | 68.69 | #187 (Qwen 3.5 27B) |
| 14 | Qwen 3 235B A22B (Thinking) | 68.29 | #313 (Qwen 3 235B A22B) |
| 15 | Qwen 3 8B (Thinking) | 67.8 | #674 (Qwen 3 8B) |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=medmarks-med-halt-reasoning-nota · How It Works · Data refreshed daily, snapshot 2026-10-20.