Video Reasoning on Maze (test)
82.1PrecisionVideoRLVR
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| VideoRLVRModel Category=RL Model, Training Epochs=1 (from SFT Epoch 5)2026.05 | 82.1 | 86.9 | 84.4 | 72.2 | |
| SFT Epoch 10Model Category=SFT Models, Training Epochs=102026.05 | 80.4 | 85.1 | 82.7 | 69 | |
| SFT Epoch 5Model Category=SFT Models, Training Epochs=52026.05 | 80.2 | 83 | 81.6 | 66.1 | |
| VBVR-Wan2.2Model Category=SFT Models2026.05 | 62.7 | 77.8 | 69.4 | 60.8 | |
| Kling V3Model Category=Proprietary2026.05 | 24.8 | 15.7 | 19.2 | 23.5 | |
| Veo 3.1Model Category=Proprietary2026.05 | 22.8 | 18.1 | 20.2 | 26 | |
| Wan-R1Model Category=SFT Models2026.05 | 20.9 | 65.6 | 31.7 | 31.9 | |
| Wan2.2-TI2V-5BModel Category=Open-Source2026.05 | 18.3 | 12.2 | 14.6 | 0 | |
| HunyuanVideoModel Category=Open-Source2026.05 | 17.3 | 11.4 | 13.8 | 2.2 | |
| Sora 2Model Category=Proprietary2026.05 | 15.8 | 17.2 | 16.5 | 3.1 | |
| CogVideoX1.5Model Category=Open-Source2026.05 | 13.3 | 10.8 | 11.9 | 0 |