Vision-Centric Evaluation on MMVP
65.2ScoreLaVer
Evaluation results
| Method | Links | |
|---|---|---|
| LaVerVisual Encoder Model=Qwen-ViT2025.12 | 65.2 | |
| BaselineVisual Encoder Model=Qwen-ViT2025.12 | 63.52 | |
| LaVerVisual Encoder Model=SigLIP22025.12 | 50.24 | |
| BaselineVisual Encoder Model=SigLIP22025.12 | 43.52 | |
| LaVerVisual Encoder Model=CLIP2025.12 | 39.36 | |
| LaVerVisual Encoder Model=AIMv22025.12 | 37.68 | |
| BaselineVisual Encoder Model=AIMv22025.12 | 35.28 | |
| LaVerVisual Encoder Model=DINOv22025.12 | 28.64 | |
| BaselineVisual Encoder Model=DINOv22025.12 | 28.48 | |
| BaselineVisual Encoder Model=CLIP2025.12 | 27.36 | |
| LaVerVisual Encoder Model=MLP + Qwen 2.52025.12 | 19.04 | |
| BaselineVisual Encoder Model=MLP + Qwen 2.52025.12 | 18.08 |