Large Vision-Language Model Evaluation on DRBench S Subset
47.22MCQ AccuracyM3ID
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| M3IDBase Model=Qwen2-VL2026.03 | 47.22 | 41.16 | 43.87 | |
| SCI5Base Model=Qwen2-VL2026.03 | 47.22 | 42.44 | 44.58 | |
| SCI7Base Model=Qwen2-VL2026.03 | 47.22 | 45.98 | 46.54 | |
| VCDBase Model=Qwen2-VL2026.03 | 46.83 | 40.84 | 43.52 | |
| TIEBase Model=Qwen2-VL2026.03 | 45.63 | 36.66 | 40.67 | |
| SCI3Base Model=Qwen2-VL2026.03 | 44.05 | 42.44 | 43.16 | |
| SCI5Base Model=LLaVA-NeXT2026.03 | 40.6 | 60.65 | 47.95 | |
| VCDBase Model=LLaVA-NeXT2026.03 | 40.5 | 56.53 | 46.38 | |
| SCI7Base Model=LLaVA-NeXT2026.03 | 40.1 | 60.65 | 47.64 | |
| M3IDBase Model=LLaVA-NeXT2026.03 | 39.9 | 56.36 | 45.94 | |
| SCI3Base Model=LLaVA-NeXT2026.03 | 39.6 | 60.31 | 47.2 | |
| LLaVA-NeXTBase Model=LLaVA-NeXT2026.03 | 39.2 | 37.63 | 38.63 | |
| TIEBase Model=LLaVA-NeXT2026.03 | 39 | 57.56 | 45.81 | |
| Qwen2-VLBase Model=Qwen2-VL2026.03 | 38.1 | 34.41 | 36.06 |