Visual Question Answering on MedXpertQA Path
80AccuracyGemini-1.5-Flash-Preview
Evaluation Results
| Method | Links | |
|---|---|---|
| Gemini-1.5-Flash-Preview2026.05 | 80 | |
| PathoSageHost Model=GPT-4o2026.05 | 71.1 | |
| Patho-AgenticRAG2026.05 | 60 | |
| GPT-4o-mini2026.05 | 46.7 | |
| Qwen2-VL-8B-Instruct2026.05 | 24.5 | |
| InternVL2-8B2026.05 | 22.2 | |
| Qwen2-VL-72B-Instruct2026.05 | 22.2 | |
| Patho-R12026.05 | 22 | |
| LLaVA-Onevision-7B2026.05 | 16.7 | |
| InternVL2.5-8B2026.05 | 10 |