Robotic Manipulation on SIMPLER (Coke Can & Drawer Tasks)
94.6Pick Coke Can Success RateCogACT + Retrieve-then-Steer
Evaluation Results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| CogACT + Retrieve-then-Steerreport_mean_std=true, seeds=32026.05 | 94.6 | — | 85.8 | — | 75.4 | — | 79.5 | — | 62.3 | |
| CogACTreport_mean_std=true, seeds=32026.05 | 91.3 | — | 83.3 | — | 71.8 | — | 75.8 | — | 56.8 | |
| CogACTProtocol=Visual Matching2026.04 | 91 | 6.1 | 78 | 6.4 | 75.9 | 13.7 | 81.6 | 8.7 | — | |
| CogACTProtocol=Variant Aggregation2026.04 | 89.2 | 6.3 | 80.8 | 6.6 | 25.8 | 13.1 | 65.3 | 8.7 | — | |
| ST-𝜋Protocol=Variant Aggregation2026.04 | 86.7 | 5.8 | 81.4 | 6.2 | 33.6 | 13.3 | 67.2 | 8.4 | — | |
| SpatialVLAProtocol=Variant Aggregation2026.04 | 85.7 | 7 | 76.1 | 7.5 | 25.9 | 15.2 | 62.6 | 9.9 | — | |
| RT-12026.05 | 85.7 | — | 44.2 | — | 73 | — | 52.4 | — | 6.5 | |
| ST-𝜋Protocol=Visual Matching2026.04 | 85.5 | 6 | 84.1 | 6 | 68.2 | 13.7 | 79.3 | 8.6 | — | |
| 𝜋0.5Protocol=Variant Aggregation2026.04 | 84.6 | 6.1 | 78.3 | 6.8 | 31.2 | 13.5 | 64.7 | 8.8 | — | |
| 𝜋0.5Protocol=Visual Matching2026.04 | 83.1 | 6.1 | 81.4 | 6.2 | 60.9 | 14.4 | 75.1 | 8.9 | — | |
| SpatialVLAProtocol=Visual Matching2026.04 | 81.7 | 6.2 | 83.3 | 6.7 | 57.4 | 14.8 | 74.1 | 9.2 | — | |
| RT-2-X2026.05 | 78.7 | — | 77.9 | — | 25 | — | 46.3 | — | 3.7 | |
| RT-1-X2026.05 | 56.7 | — | 31.7 | — | 59.7 | — | 42.4 | — | 21.3 | |
| OpenVLAPruning Ratio=0%2025.11 | 52.8 | — | 54 | — | 49.5 | — | 52.1 | — | — | |
| VLA-PrunerPruning Ratio=75%2025.11 | 50.1 | — | 52.4 | — | 48.8 | — | 50.4 | — | — | |
| OpenVLAProtocol=Variant Aggregation2026.04 | 48.2 | 12.6 | 43.5 | 11.7 | 15.1 | 17.6 | 35.6 | 14 | — | |
| FastVPruning Ratio=75%2025.11 | 41.9 | — | 38.7 | — | 33.7 | — | 38.1 | — | — | |
| VLA-CachePruning Ratio=75%2025.11 | 40.6 | — | 41.2 | — | 38.8 | — | 40.2 | — | — | |
| OpenVLA2026.05 | 18 | — | 56.3 | — | 63 | — | 34.3 | — | 0 | |
| Octo-Base2026.05 | 17 | — | 4.2 | — | 22.7 | — | 11 | — | 0 | |
| OpenVLAProtocol=Visual Matching2026.04 | 14.7 | 15.2 | 49.2 | 11.8 | 39.8 | 15.5 | 34.6 | 14.2 | — | |
| OctoProtocol=Visual Matching2026.04 | 14.7 | 11.8 | 4.2 | 10.3 | 24.1 | 18.4 | 14.3 | 13.5 | — | |
| OctoProtocol=Variant Aggregation2026.04 | 0.2 | 15.8 | 2.8 | 13.6 | 0.8 | 25.8 | 1.3 | 18.4 | — |