Visual Reasoning on CVBench
76.9ScoreBaseline
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| BaselineCheckpoint=VL-Rethinker [18], Pruning rate (ρ)=0%, Rollouts (K)=12026.06 | 76.9 | 100 | |
| AVISCheckpoint=VL-Rethinker [18], Pruning rate (ρ)=Adaptive, Rollouts (K)=Adaptive2026.06 | 76.1 | 42 | |
| AVISCheckpoint=OpenVLThinker [75], Pruning rate (ρ)=Adaptive, Rollouts (K)=Adaptive2026.06 | 74.8 | 39 | |
| AVISCheckpoint=Vision-R1 [5], Pruning rate (ρ)=Adaptive, Rollouts (K)=Adaptive2026.06 | 74.4 | 43 | |
| FixedCheckpoint=Qwen 2.5 VL [1], Pruning rate (ρ)=75%, Rollouts (K)=52026.06 | 74.3 | 55 | |
| FixedCheckpoint=OpenVLThinker [75], Pruning rate (ρ)=75%, Rollouts (K)=52026.06 | 74.2 | 50 | |
| FixedCheckpoint=VL-Rethinker [18], Pruning rate (ρ)=75%, Rollouts (K)=52026.06 | 73.6 | 64 | |
| FixedCheckpoint=Vision-R1 [5], Pruning rate (ρ)=75%, Rollouts (K)=52026.06 | 73.4 | 56 | |
| AVISCheckpoint=Qwen 2.5 VL [1], Pruning rate (ρ)=Adaptive, Rollouts (K)=Adaptive2026.06 | 73.1 | 43 | |
| BaselineCheckpoint=OpenVLThinker [75], Pruning rate (ρ)=0%, Rollouts (K)=12026.06 | 72.9 | 100 | |
| BaselineCheckpoint=Vision-R1 [5], Pruning rate (ρ)=0%, Rollouts (K)=12026.06 | 72.7 | 100 | |
| BaselineCheckpoint=Qwen 2.5 VL [1], Pruning rate (ρ)=0%, Rollouts (K)=12026.06 | 70.6 | 100 |