Visual Planning Quality Evaluation on RoboTwin 210 manipulation prompts 2.0 (test)
91.4Kinematic Success RateEVA
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| EVARL alignment=true2026.03 | 91.4 | 86.2 | 89.5 | 83.8 | |
| EVARL alignment=false2026.03 | 70.5 | 83.3 | 90.5 | 68.1 | |
| Vidar (Wan2.2)Method description=Vidar (Wan2.2)2026.03 | 67.6 | 66.7 | 87.6 | 62.9 |