DermoBench - OOD Diagnosis - Derm7pt: leaderboard
Metric: Accuracy (%; 4-way MCQA in the source dataset's own label space, out-of-distribution). Source: arxiv.org. Saturation forecast: Around March 2027. 18 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3 VL 8B Instruct | 59.1 |
| 2 | Qwen 2.5 VL 72B Instruct | 58.36 |
| 3 | QVQ-72B-Preview | 56.92 |
| 4 | GPT-4o Mini | 56.48 |
| 5 | Gemini 2.5 Flash | 53.96 |
| 6 | Claude Sonnet 4.5 (Thinking) | 50.4 |
| 7 | GLM-4.5V | 43.08 |
| 8 | Llama 3.2 90B | 37.14 |
| 9 | Nemotron Nano 12B v2 VL | 36.84 |
| 10 | Lingshu-7B | 33.88 |
Interactive version: theaggregate.ai/benchmark?slug=dermobench-ood-diagnosis-derm7pt · How It Works · Data refreshed daily, snapshot 2026-09-25.