Spatial Reasoning on V-Star
92.5Average ScoreToolsRL
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| ToolsRLSFT=false, RL=true2026.04 | 92.5 | 95.6 | 89.4 | |
| Simple o3SFT=true, RL=false2026.04 | 90.4 | — | — | |
| DeepEyesSFT=false, RL=true2026.04 | 89.8 | 91.3 | 88.2 | |
| ZoomEyeSFT=false, RL=false2026.04 | 89.7 | 93.9 | 85.5 | |
| Mini-o3SFT=true, RL=true2026.04 | 88.2 | — | — | |
| Pixel-ReasonerSFT=true, RL=true2026.04 | 86.3 | — | — | |
| Qwen2.5-VLSFT=false, RL=false2026.04 | 75.9 | 78.2 | 73.6 |