Medical Visual Question Answering on M3D
92.8Plane ScoreDART
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| DART2025.12 | 92.8 | 40.7 | 47.6 | 55.9 | 51.8 | 57.8 | |
| QwenVL2025.12 | 91.2 | 33.5 | 49.7 | 60 | 47.7 | 56.4 | |
| GPT-4o2025.12 | 83.3 | 42.7 | 50 | 41.3 | 41.3 | 51.7 | |
| Debate with Consensusstrategy=Consensus2025.12 | 81.2 | 32 | 43.1 | 60.9 | 51.9 | 53.7 | |
| MiniCPM-o2025.12 | 79.8 | 28.4 | 53.9 | 49.2 | 37.8 | 49.8 | |
| Debate with Judgestrategy=Judge2025.12 | 79.5 | 32 | 45.6 | 61.2 | 51.7 | 53.4 | |
| MEXA2025.12 | 65 | 48.1 | 60.9 | 44.8 | 48 | 53.3 | |
| Ovis22025.12 | 59.6 | 34 | 39.8 | 55.4 | 43 | 46.4 | |
| CREMA2025.12 | 14.9 | 26.7 | 15.9 | 17.3 | 13 | 17.2 |