Vision-Language Performance on MM-Vet
49.13Utility ScoreVanilla
Evaluation Results
| Method | Links | |
|---|---|---|
| VanillaBase Model=Qwen2-VL2026.03 | 49.13 | |
| NullSteerBase Model=Qwen2-VL2026.03 | 49.02 | |
| ASTRABase Model=Qwen2-VL2026.03 | 48.66 | |
| VanillaBase Model=LLaVA-v1.52026.03 | 32.62 | |
| NullSteerBase Model=LLaVA-v1.52026.03 | 31.2 | |
| ASTRABase Model=LLaVA-v1.52026.03 | 30.55 | |
| NullSteerBase Model=MiniGPT-42026.03 | 21.05 | |
| ASTRABase Model=MiniGPT-42026.03 | 20.62 | |
| VanillaBase Model=MiniGPT-42026.03 | 19.4 |