Medical MM Leaderboard - VQA-RAD: leaderboard
The VQA-RAD slice of the Medical Multimodal Leaderboard — clinician-authored questions about radiology images; accuracy (%).
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Around August 2027. 22 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | InternVL3-38B | 72.7 | #405 |
| 2 | MedGemma-4B-IT | 72.5 | #853 |
| 3 | Gemini 2.5 Flash | 68.5 | #245 |
| 4 | Claude Sonnet 4 | 67.6 | #199 |
| 5 | InternVL3-14B | 66.1 | #505 |
| 6 | GPT-4.1 | 65 | #248 |
| 7 | InternVL3-8B | 61.4 | #620 |
| 8 | Qwen 2.5 VL 7B | 50.6 | #643 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=medical-mm-leaderboard-vqa-rad · How It Works · Data refreshed daily, snapshot 2026-10-20.