Visual Language Model Evaluation on MMVet V2
52.6MMVet V2 ScorePOW3R dynamic
Evaluation Results
| Method | Links | |
|---|---|---|
| POW3R dynamicBase policy=Qwen3-VL-8B2026.05 | 52.6 | |
| POW3R dynamicBase policy=Qwen3-VL-4B2026.05 | 52.2 | |
| Category-balancedBase policy=Qwen3-VL-4B2026.05 | 51.8 | |
| Category-balancedBase policy=Qwen3-VL-8B2026.05 | 51.8 | |
| Static scalarBase policy=Qwen3-VL-4B2026.05 | 51 | |
| Static scalarBase policy=Qwen3-VL-8B2026.05 | 51 | |
| BaseBase policy=Qwen3-VL-4B2026.05 | 49.9 | |
| BinaryBase policy=Qwen3-VL-4B2026.05 | 49.8 | |
| BaseBase policy=Qwen3-VL-8B2026.05 | 49.7 | |
| BinaryBase policy=Qwen3-VL-8B2026.05 | 49.5 | |
| POW3R dynamicBase policy=Gemma3-4B2026.05 | 40.8 | |
| Category-balancedBase policy=Gemma3-4B2026.05 | 40.2 | |
| Static scalarBase policy=Gemma3-4B2026.05 | 39.5 | |
| BaseBase policy=Gemma3-4B2026.05 | 38.9 | |
| BinaryBase policy=Gemma3-4B2026.05 | 38.8 |