Rigid-object placement on Place Coffee
90Coarse Success RateMSARNN
Evaluation results
| Method | Links | ||
|---|---|---|---|
| MSARNNInput Image=Stereo, Visual Encoder Backbone=MSA2026.05 | 90 | 72 | |
| ACTInput Image=Stereo, Visual Encoder Backbone=ResNet2026.05 | 62 | 48 | |
| DPInput Image=Stereo, Visual Encoder Backbone=ResNet2026.05 | 46 | 42 | |
| π0Input Image=Stereo, Visual Encoder Backbone=ViT, Parameters=3.5B2026.05 | 40 | 16 | |
| MSARNNInput Image=Mono, Visual Encoder Backbone=MSA2026.05 | 40 | 18 | |
| SmolVLAInput Image=Stereo, Visual Encoder Backbone=ViT, Parameters=0.45B2026.05 | 22 | 18 | |
| MSARNNInput Image=Stereo, Visual Encoder Backbone=SA2026.05 | 22 | 8 |