Multiple-choice Question Answering on PathMMU EduContent (test-all)
73.5AccuracyPathChat+
Evaluation results
| Method | Links | ||
|---|---|---|---|
| PathChat+2025.06 | 73.5 | 71.6 | |
| Gemini 2.5 pro2025.06 | 69.3 | 67.2 | |
| GPT-52025.06 | 68.9 | 66.9 | |
| GPT-5-mini2025.06 | 66.9 | 64.6 | |
| Claude Sonnet 42025.06 | 63.3 | 61.1 | |
| PathChat 12025.06 | 63.1 | 60.9 | |
| HuatuoGPT-Vision2025.06 | 57 | 54.8 | |
| Qwen3-VL2025.06 | 56.1 | 53.9 | |
| LLaVA-OneVision2025.06 | 51.2 | 48.9 | |
| Llama-3.22025.06 | 50.7 | 48.8 | |
| Quilt-LLaVA2025.06 | 43.7 | 41.3 | |
| PA-LLaVA2025.06 | 42.1 | 39.9 |