Visual Question Answering on PathMMU All-tiny (test)
77.1AccuracyPathChat+
Evaluation results
| Method | Links | ||
|---|---|---|---|
| PathChat+2025.06 | 77.1 | 74.7 | |
| GPT-52025.06 | 71.9 | 69.2 | |
| Expert performance2025.06 | 71.8 | — | |
| Gemini 2.5 pro2025.06 | 71.4 | 69 | |
| GPT-5-mini2025.06 | 68.3 | 65.6 | |
| PathChat 12025.06 | 64 | 61.3 | |
| Claude Sonnet 42025.06 | 61 | 58.1 | |
| HuatuoGPT-Vision2025.06 | 58.7 | 56.1 | |
| Qwen3-VL2025.06 | 57.1 | 54.3 | |
| Llama-3.22025.06 | 53 | 50.1 | |
| LLaVA-OneVision2025.06 | 48.9 | 45.9 | |
| Quilt-LLaVA2025.06 | 44.1 | 41.3 | |
| PA-LLaVA2025.06 | 42.2 | 39.6 |