Mobile manipulation on Mobile Manipulation Task Suite
85Average Success RateMSARNN
Evaluation Results
| Method | Links | |
|---|---|---|
| MSARNNInput Image=Stereo, Visual Encoder Backbone=MSA2026.05 | 85 | |
| ACTInput Image=Stereo, Visual Encoder Backbone=ResNet2026.05 | 46 | |
| MSARNNInput Image=Stereo, Visual Encoder Backbone=SA2026.05 | 37.5 | |
| MSARNNInput Image=Mono, Visual Encoder Backbone=MSA2026.05 | 33 | |
| π0Input Image=Stereo, Visual Encoder Backbone=ViT, Parameters=3.5B2026.05 | 29 | |
| DPInput Image=Stereo, Visual Encoder Backbone=ResNet2026.05 | 28.5 | |
| SmolVLAInput Image=Stereo, Visual Encoder Backbone=ViT, Parameters=0.45B2026.05 | 12.5 |