MedXpertQA MM (Natomy): leaderboard
Accuracy on the MedXpertQA multimodal subset - 2,000 expert-level medical exam questions with real clinical images and five answer choices - run zero-shot with chain-of-thought reasoning.
Metric: Accuracy (%). Source: www.natomy.com. Saturation forecast: Around February 2027. 14 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 4 Argon | 90.5 | #9 |
| 2 | GPT-6 Astra | 87.4 | #6 |
| 3 | Gemini 3.8 Flash | 86.55 | #30 |
| 4 | Claude Opus 5.5 (High) | 86.55 | #3 (Claude Opus 5.5) |
| 5 | Gemini 3.1 Pro (Preview) | 82.35 | #64 |
| 6 | GPT-5.6 Sol | 77.7 | #22 |
| 7 | Claude Opus 5 | 76.25 | #16 |
| 8 | GPT-5.6 Terra | 72 | #40 |
| 9 | GPT-5.6 Luna (High) | 69 | #79 (GPT-5.6 Luna) |
| 10 | GPT-5.6 Luna (Medium) | 67.25 | #79 (GPT-5.6 Luna) |
| 11 | Claude Sonnet 5 | 58.85 | #65 |
| 12 | GPT-4o | 42.8 | #357 |
| 13 | Gemini 2.0 Flash | 37.2 | #351 |
| 14 | Claude 3.5 Sonnet | 33.2 | #362 |
No result here: #7 GPT-6.1 Sol, #10 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=medxpertqa-mm-natomy · How It Works · Data refreshed daily, snapshot 2026-11-05.