Robotic Manipulation on UT Austin MUTEX
77.26Success Rate (%)VISUALTHINK-VLA
Evaluation results
| Method | Links | ||
|---|---|---|---|
| VISUALTHINK-VLAReasoning Strategy=OpenVLA-family re-evaluation2026.05 | 77.26 | 0.451 | |
| FULLSOFTReasoning Strategy=OpenVLA-family re-evaluation2026.05 | 77.1 | 0.551 | |
| BASEVLAReasoning Strategy=OpenVLA-family re-evaluation2026.05 | 41.09 | 0.349 | |
| DeepThinkVLA-RLReasoning Strategy=Reinforced or hybrid thinking-action decoding2026.05 | 26.09 | 0.772 | |
| ECoTReasoning Strategy=Textual CoT / autoregressive reasoning before action2026.05 | 20.32 | 6.255 | |
| InternVLA-M1Reasoning Strategy=Spatial or image-grounded thinking2026.05 | 12.37 | 1.144 | |
| TraceVLAReasoning Strategy=Spatial or image-grounded thinking2026.05 | 1.1 | 0.403 | |
| SpatialVLAReasoning Strategy=Spatial or image-grounded thinking2026.05 | 0.65 | 0.544 |