| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| LIBERO | GR00T-N1.6-ft + World2Act | S (Spatial) Rate99.5 | 73 | 24d ago | |
| LIBERO standard 4-suite protocol | π0.5 | Spatial Achievement Rate98.8 | 10 | 26d ago | |
| Franka Kitchen online evaluation in simulation | CroBo | Knob Turn Success65.6 | 8 | 4mo ago | |
| RLBench | ToBo | Button Success Rate41.2 | 8 | 4mo ago | |
| DROID Franka Panda | KT + Mix-Mask + FiLM | Average Success Rate47.4 | 7 | 4mo ago | |
| DeepMind Control Suite (DMC) online evaluation in simulation | CroBo | Walker/Stand Score92 | 5 | 4mo ago | |
| RLBench Unimodal (test) | MiDiGaP | Open Drawer Steps61 | 3 | 1mo ago | |
| Real-World 8 Tasks | Dopamine-RL | Success Rate (%)95.2 | 3 | 5mo ago | |
| Simulation 10 Tasks | Dopamine-RL | SR0.81 | 3 | 5mo ago | |
| DROID | Cosmos3-Nano | Score39.7 | 2 | 1mo ago |