Multiple-choice Visual Question Answering on PathMMU PubMed (test-all)
75.3AccuracyPathChat+
Evaluation results
| Method | Links | ||
|---|---|---|---|
| PathChat+2025.06 | 75.3 | 73.8 | |
| Gemini 2.5 pro2025.06 | 72.5 | 70.9 | |
| GPT-52025.06 | 71.3 | 69.6 | |
| GPT-5-mini2025.06 | 68.8 | 67.1 | |
| Claude Sonnet 42025.06 | 64.1 | 62.3 | |
| PathChat 12025.06 | 63.6 | 62 | |
| HuatuoGPT-Vision2025.06 | 63.5 | 61.9 | |
| Qwen3-VL2025.06 | 57.9 | 56.1 | |
| Llama-3.22025.06 | 53.5 | 51.7 | |
| LLaVA-OneVision2025.06 | 48.9 | 47.1 | |
| Quilt-LLaVA2025.06 | 41.6 | 39.9 | |
| PA-LLaVA2025.06 | 40.1 | 38.4 |