Visual Question Answering on PMC-MI-Bench single-image
15.4BLEU@4M3LLM-8B
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| M3LLM-8Bparameter_count=8B2025.11 | 15.4 | 38.4 | 65.8 | 82.5 | |
| LLaVA-Med-7Bparameter_count=7B2025.11 | 11.6 | 34.5 | 67.2 | 79.4 | |
| Lingshu-7Bparameter_count=7B2025.11 | 10 | 33.8 | 66.3 | 80 | |
| HealthGPT-14Bparameter_count=14B2025.11 | 9.8 | 34 | 67.1 | 79.8 | |
| HuatuoGPT-Vision-7Bparameter_count=7B2025.11 | 9.1 | 31.7 | 65 | 79.2 | |
| InternVL3-8Bparameter_count=8B2025.11 | 6.8 | 29.5 | 58.7 | 78.6 | |
| QWen2.5-VL-7Bparameter_count=7B2025.11 | 3.4 | 23.5 | 55.5 | 73.8 | |
| LLaVA-NeXT-7Bparameter_count=7B2025.11 | 2.7 | 20.6 | 66 | 67.4 | |
| LLaVA-7Bparameter_count=7B2025.11 | 2.3 | 22 | 55.3 | 70.1 | |
| MedGemma-27Bparameter_count=27B2025.11 | 2.3 | 19 | 51.9 | 73.4 |