Large Vision-Language Model Evaluation on DRBench B
27.04MCQ ScoreSCI7
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| SCI7Base Model=Qwen2-VL2026.03 | 27.04 | 29.66 | 27.65 | |
| SCI5Base Model=Qwen2-VL2026.03 | 24.91 | 25.69 | 25.09 | |
| SCI7Base Model=LLaVA-NeXT2026.03 | 24.86 | 38.26 | 27.01 | |
| SCI5Base Model=LLaVA-NeXT2026.03 | 23.81 | 37.97 | 26.08 | |
| SCI3Base Model=Qwen2-VL2026.03 | 21.67 | 26.3 | 22.74 | |
| SCI3Base Model=LLaVA-NeXT2026.03 | 21.22 | 35.36 | 23.48 | |
| M3IDBase Model=Qwen2-VL2026.03 | 19.81 | 21.71 | 20.26 | |
| M3IDBase Model=LLaVA-NeXT2026.03 | 16.91 | 25.22 | 18.24 | |
| TIEBase Model=Qwen2-VL2026.03 | 16.2 | 16.82 | 16.35 | |
| VCDBase Model=Qwen2-VL2026.03 | 15.74 | 21.71 | 17.13 | |
| TIEBase Model=LLaVA-NeXT2026.03 | 12.98 | 23.48 | 14.66 | |
| VCDBase Model=LLaVA-NeXT2026.03 | 12.65 | 25.51 | 14.71 | |
| Qwen2-VLBase Model=Qwen2-VL2026.03 | 5.37 | 8.56 | 6.11 | |
| LLaVA-NeXTBase Model=LLaVA-NeXT2026.03 | 0 | 0 | 0 |