CareQA-Vision - Open-Ended - Medicine: leaderboard

Metric: Judge score (%; 108 open-ended image-based questions from MIR medicine exams, majority of three LLM judges scoring 0, 0.5 or 1). Source: arxiv.org. Saturation forecast: Around December 2026. 18 models tracked.

Top models

#ModelScore
1GLM-4.5V68.06
2Qwen 3 VL 30B A3B Instruct49.07
3Qwen 2 VL 72B Instruct49.07
4HuluMed-7B48.61
5Qwen 3 VL 8B Instruct41.2
6Lingshu-7B34.72
7Qwen 2 VL 7B Instruct29.17

Interactive version: theaggregate.ai/benchmark?slug=careqa-vision-open-ended-medicine · How It Works · Data refreshed daily, snapshot 2026-09-25.