CareQA-Vision - MCQ: leaderboard
Image-based multiple-choice questions from the 2020-2024 Spanish MIR (medicine) and EIR (nursing) residency entrance exams, translated to English; CareQA-Vision (Guasch-Martí et al., Barcelona Supercomputing Center, 2026). Accuracy in percent over 165 questions, scored by exact match under greedy decoding; higher is better.
Metric: Accuracy (%; 165 image-based multiple-choice questions from MIR and EIR exams, exact match). Source: arxiv.org. Saturation forecast: Around February 2027. 19 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.2 | 80.61 | #172 |
| 2 | GLM-4.5V | 72.73 | #739 |
| 3 | Qwen 2 VL 72B Instruct | 72.73 | #1403 |
| 4 | Qwen 3 VL 30B A3B Instruct | 71.52 | #870 |
| 5 | Qwen 3 VL 8B Instruct | 58.79 | #982 |
| 6 | Qwen 2 VL 7B Instruct | 51.52 | #2569 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=careqa-vision-mcq · How It Works · Data refreshed daily, snapshot 2026-11-12.