Med-R2 - Joint QA: leaderboard
Accuracy (%) of medical vision-language models on the Med-R2 (2026) joint qa task, answering all reasoning levels for one image in a single prompt, with equal numbers of positive, negative and neutral adversarial question forms; higher is better.
Metric: Accuracy (%) on the Med-R2 Joint QA task (answering all reasoning levels for one image in a single prompt), zero-shot with system prompts; each task mixes equal numbers of positive, negative and neutral adversarial question forms; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 14 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Setting | Score | Overall rank |
|---|---|---|---|---|
| 1 | GPT-5.2 | Thinking | 90.2 | #185 |
| 2 | Qwen 3 VL 8B | 75.3 | #917 | |
| 3 | InternVL3.5-8B | 73.8 | #1604 | |
| 4 | Qwen 3 VL 235B A22B | 72.5 | #485 | |
| 5 | GPT-4o | 65.5 | #751 | |
| 6 | Hulu-Med-14B | 63.1 | #1607 | |
| 7 | Lingshu-7B | 60.2 | #1635 | |
| 8 | Hulu-Med-4B | 58.4 | #2392 | |
| 9 | Qwen 2.5 VL 7B Instruct | 57.3 | #1902 | |
| 10 | HuaTuoGPT-Vision-34B | 54.8 | #1828 | |
| 11 | HuatuoGPT-Vision-7B | 51 | #2162 | |
| 12 | HuluMed-7B | 48.9 | #2049 | |
| 13 | MedGemma-4B | 45.3 | #2557 | |
| 14 | Janus-Pro-7B | 32.7 | #6857 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=med-r2-joint-qa · How It Works · Data refreshed daily, snapshot 2026-11-21.