Robot Manipulation on LIBERO-Plus Zero-shot
82.8Camera Scoreπ0.5 + GridS32
Evaluation Results
| Method | Links | ||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| π0.5 + GridS32evaluation_mode=Zero-shot OOD, pruning_method=GridS322026.05 | 82.8 | 61.8 | 80.8 | 97.6 | 98 | 97.9 | 87.6 | 85.7 | — | 93.9 | 92.8 | 90.3 | 83.4 | 72.4 | |
| π0.5 Baselineevaluation_mode=Zero-shot OOD, pruning_method=None (Baseline)2026.05 | 81.3 | 73.6 | 92.4 | 99 | 99.6 | 95.5 | 90.1 | 89.3 | — | 97.3 | 96.7 | 94.7 | 86.1 | 75.9 | |
| OA-WAMModel Category=World-Action Models, Evaluation Mode=zero-shot, Training Data=standard LIBERO only2026.05 | 80.5 | 89.6 | 85.3 | 96.5 | 95.9 | 75.6 | 82.8 | 83.9 | 84.3 | — | — | — | — | — | |
| Cosmos-PolicyModel Category=World-Action Models, Evaluation Mode=zero-shot, Training Data=standard LIBERO only2026.05 | 75.8 | 63.3 | 81.7 | 96.5 | 88.9 | 92.7 | 82.2 | 82.2 | 73.8 | — | — | — | — | — | |
| π0.5Model Category=Vision-Language-Action Models, Evaluation Mode=zero-shot, Training Data=standard LIBERO only2026.05 | 75.4 | 77.5 | 85.6 | 96.9 | 94.6 | 89.7 | 85.7 | 85.7 | 79.5 | — | — | — | — | — | |
| HoloBrain-0-QWSize=1.1B2026.02 | 66.3 | 49 | 65.9 | 94.9 | 93.3 | 73.1 | 78.2 | 72.6 | — | — | — | — | — | — | |
| HoloBrain-0-GDSize=0.2B2026.02 | 65.5 | 58.2 | 78.7 | 88.1 | 90.3 | 66.9 | 79.5 | 74 | — | — | — | — | — | — | |
| HoloBrain-0Model Category=World-Action Models, Evaluation Mode=zero-shot, Training Data=standard LIBERO only2026.05 | 65.5 | 58.2 | 78.7 | 88.1 | 90.3 | 66.9 | 79.5 | 74 | 67.7 | — | — | — | — | — | |
| pi0-FastEvaluation protocol=Zero-shot, Training dataset=Standard LIBERO2026.02 | 65.1 | 21.6 | 61 | 73.2 | 73.2 | 74.4 | 68.8 | 61.6 | — | — | — | — | — | — | |
| π0+FASTSize=3B2026.02 | 65.1 | 21.6 | 61 | 73.2 | 73.2 | 74.4 | 68.8 | 61.6 | — | — | — | — | — | — | |
| π0 -Fastzero-shot=true2026.03 | 65.1 | 21.6 | 61 | 73.2 | 73.2 | 74.4 | 68.8 | 61.6 | — | — | — | — | — | — | |
| π0-Fasttraining=standard LIBERO, protocol=zero-shot2026.07 | 65.1 | 21.6 | 61 | 73.2 | 73.2 | 74.4 | 68.8 | 61.6 | — | — | — | — | — | — | |
| LaMPzero-shot=true2026.03 | 64.5 | 69.6 | 88.2 | 95.3 | 97.4 | 76.9 | 73.8 | 79.3 | — | — | — | — | — | — | |
| VLA-JEPAModel Category=World-Action Models, Evaluation Mode=zero-shot, Training Data=standard LIBERO only2026.05 | 64.2 | 67.7 | 88.1 | 91.8 | 93.4 | 65.8 | 83.9 | 79.5 | 71.9 | — | — | — | — | — | |
| GE-ActModel Category=World-Action Models, Evaluation Mode=zero-shot, Training Data=standard LIBERO only2026.05 | 60.7 | 77 | 77.4 | 95.8 | 86 | 90.9 | 80.2 | 80.3 | 72.6 | — | — | — | — | — | |
| ABot-M0Evaluation protocol=Zero-shot, Training dataset=Standard LIBERO2026.02 | 60.4 | 67.9 | 86.4 | 96.2 | 91.6 | 86.4 | 82.6 | 80.5 | — | — | — | — | — | — | |
| ABot-M0Model Category=Vision-Language-Action Models, Evaluation Mode=zero-shot, Training Data=standard LIBERO only2026.05 | 60.4 | 67.9 | 86.4 | 96.2 | 91.6 | 86.4 | 82.6 | 80.5 | 70.3 | — | — | — | — | — | |
| OpenVLA-OFTEvaluation protocol=Zero-shot, Training dataset=Standard LIBERO2026.02 | 56.4 | 31.9 | 79.5 | 88.7 | 93.3 | 75.8 | 74.2 | 69.6 | — | — | — | — | — | — | |
| OpenVLA-OFTSize=7B2026.02 | 56.4 | 31.9 | 79.5 | 88.7 | 93.3 | 75.8 | 74.2 | 69.6 | — | — | — | — | — | — | |
| OpenVLA-OFTzero-shot=true2026.03 | 56.4 | 31.9 | 79.5 | 88.7 | 93.3 | 75.8 | 74.2 | 69.6 | — | — | — | — | — | — | |
| OpenVLA-OFTModel Category=Vision-Language-Action Models, Evaluation Mode=zero-shot, Training Data=standard LIBERO only2026.05 | 56.4 | 31.9 | 79.5 | 88.7 | 93.3 | 75.8 | 74.2 | 69.6 | 54.2 | — | — | — | — | — | |
| OpenVLA-OFTtraining=standard LIBERO, protocol=zero-shot2026.07 | 56.4 | 31.9 | 79.5 | 88.7 | 93.3 | 75.8 | 74.2 | 69.6 | — | — | — | — | — | — | |
| Openvla-OFT_mEvaluation protocol=Zero-shot, Training dataset=Standard LIBERO2026.02 | 55.6 | 21.7 | 81 | 92.7 | 91 | 78.6 | 68.7 | 67.9 | — | — | — | — | — | — | |
| OpenVLA-OFT-mtraining=standard LIBERO, protocol=zero-shot2026.07 | 55.6 | 21.7 | 81 | 92.7 | 91 | 78.6 | 68.7 | 67.9 | — | — | — | — | — | — | |
| AVA-VLAModel Category=Vision-Language-Action Models, Evaluation Mode=zero-shot, Training Data=standard LIBERO only2026.05 | 55.5 | 25.9 | 85.6 | 95.5 | 88.9 | 78 | 74.1 | 70.1 | 51.8 | — | — | — | — | — | |
| RIPT-VLAEvaluation protocol=Zero-shot, Training dataset=Standard LIBERO2026.02 | 55.2 | 31.2 | 77.6 | 88.4 | 91.6 | 73.5 | 74.2 | 68.4 | — | — | — | — | — | — | |
| RIPT-VLAtraining=standard LIBERO, protocol=zero-shot2026.07 | 55.2 | 31.2 | 77.6 | 88.4 | 91.6 | 73.5 | 74.2 | 68.4 | — | — | — | — | — | — | |
| LaMP (w/o motion)zero-shot=true2026.03 | 46.7 | 56 | 82.5 | 95.3 | 95.4 | 69.3 | 71 | 71.6 | — | — | — | — | — | — | |
| MindLPItraining=standard LIBERO, protocol=zero-shot2026.07 | 46.2 | 61.7 | 84.6 | 93.4 | 95.2 | 58.4 | 82.2 | 72.3 | — | — | — | — | — | — | |
| MindPI (Frozen VLM)backbone=Frozen VLM, training=standard LIBERO, protocol=zero-shot2026.07 | 42.3 | 58.6 | 86.6 | 94.7 | 96.3 | 81.9 | 78 | 74.9 | — | — | — | — | — | — | |
| MindPI (Full PT)variant=Full Pre-training, training=standard LIBERO, protocol=zero-shot2026.07 | 42 | 57.8 | 75.4 | 91.5 | 93.2 | 57.4 | 86.7 | 68.8 | — | — | — | — | — | — | |
| MindWPItraining=standard LIBERO, protocol=zero-shot2026.07 | 36.7 | 73.8 | 86.2 | 96.5 | 91.6 | 54.2 | 84.8 | 72.6 | — | — | — | — | — | — | |
| MindLWPItraining=standard LIBERO, protocol=zero-shot2026.07 | 35.1 | 74.2 | 93 | 97.1 | 96.7 | 58.9 | 84.5 | 74.8 | — | — | — | — | — | — | |
| MindWPI w/o PTvariant=without pre-training, training=standard LIBERO, protocol=zero-shot2026.07 | 32.7 | 48.7 | 77 | 90.7 | 93.2 | 64 | 73.6 | 66.1 | — | — | — | — | — | — | |
| MindPI w/o PTvariant=without pre-training, training=standard LIBERO, protocol=zero-shot2026.07 | 26.1 | 32 | 71 | 88 | 94.4 | 58.3 | 70.3 | 59.9 | — | — | — | — | — | — | |
| X-VLAModel Category=Vision-Language-Action Models, Evaluation Mode=zero-shot, Training Data=standard LIBERO only2026.05 | 23.4 | 89.7 | 75.7 | 88.2 | 96 | 62.7 | 71.8 | 71.4 | 61.6 | — | — | — | — | — | |
| X-VLASize=0.9B2026.02 | 22.2 | 87.8 | 73.1 | 88.2 | 95.3 | 61.8 | 70.7 | 69.7 | — | — | — | — | — | — | |
| pi0Evaluation protocol=Zero-shot, Training dataset=Standard LIBERO2026.02 | 13.8 | 6 | 58.8 | 85 | 81.4 | 79 | 68.9 | 53.6 | — | — | — | — | — | — | |
| π0Size=3B2026.02 | 13.8 | 6 | 58.8 | 85 | 81.4 | 79 | 68.9 | 53.6 | — | — | — | — | — | — | |
| π0zero-shot=true2026.03 | 13.8 | 6 | 58.8 | 85 | 81.4 | 79 | 68.9 | 53.6 | — | — | — | — | — | — | |
| π0Model Category=Vision-Language-Action Models, Evaluation Mode=zero-shot, Training Data=standard LIBERO only2026.05 | 13.8 | 6 | 58.8 | 85 | 81.4 | 79 | 68.9 | 53.6 | 29.6 | — | — | — | — | — | |
| π0training=standard LIBERO, protocol=zero-shot2026.07 | 13.8 | 6 | 58.8 | 85 | 81.4 | 79 | 68.9 | 53.6 | — | — | — | — | — | — | |
| OpenVLA-OFT_wEvaluation protocol=Zero-shot, Training dataset=Standard LIBERO2026.02 | 10.4 | 38.7 | 70.5 | 76.8 | 93.6 | 49.9 | 69.9 | 55.8 | — | — | — | — | — | — | |
| OpenVLA-OFT-wtraining=standard LIBERO, protocol=zero-shot2026.07 | 10.4 | 38.7 | 70.5 | 76.8 | 93.6 | 49.9 | 69.9 | 55.8 | — | — | — | — | — | — | |
| NORAEvaluation protocol=Zero-shot, Training dataset=Standard LIBERO2026.02 | 2.2 | 37 | 65.1 | 45.7 | 58.6 | 12.8 | 62.1 | 39 | — | — | — | — | — | — | |
| NORAtraining=standard LIBERO, protocol=zero-shot2026.07 | 2.2 | 37 | 65.1 | 45.7 | 58.6 | 12.8 | 62.1 | 39 | — | — | — | — | — | — | |
| UniVLAEvaluation protocol=Zero-shot, Training dataset=Standard LIBERO2026.02 | 1.8 | 46.2 | 69.6 | 69 | 81 | 21.2 | 31.9 | 42.9 | — | — | — | — | — | — | |
| UniVLASize=7B2026.02 | 1.8 | 46.2 | 69.6 | 69 | 81 | 21.2 | 31.9 | 42.9 | — | — | — | — | — | — | |
| UniVLAzero-shot=true2026.03 | 1.8 | 46.2 | 69.6 | 69 | 81 | 21.2 | 31.9 | 42.9 | — | — | — | — | — | — | |
| UniVLAtraining=standard LIBERO, protocol=zero-shot2026.07 | 1.8 | 46.2 | 69.6 | 69 | 81 | 21.2 | 31.9 | 42.9 | — | — | — | — | — | — | |
| OpenVLAEvaluation protocol=Zero-shot, Training dataset=Standard LIBERO2026.02 | 0.8 | 3.5 | 23 | 8.1 | 34.8 | 15.2 | 28.5 | 15.6 | — | — | — | — | — | — | |
| OpenVLASize=7B2026.02 | 0.8 | 3.5 | 23 | 8.1 | 34.8 | 15.2 | 28.5 | 15.6 | — | — | — | — | — | — | |
| OpenVLAzero-shot=true2026.03 | 0.8 | 3.5 | 23 | 8.1 | 34.8 | 15.2 | 28.5 | 15.6 | — | — | — | — | — | — | |
| OpenVLAtraining=standard LIBERO, protocol=zero-shot2026.07 | 0.8 | 3.5 | 23 | 8.1 | 34.8 | 15.2 | 28.5 | 15.6 | — | — | — | — | — | — | |
| WorldVLAEvaluation protocol=Zero-shot, Training dataset=Standard LIBERO2026.02 | 0.1 | 27.9 | 41.6 | 43.7 | 17.1 | 10.9 | 38 | 25 | — | — | — | — | — | — | |
| WorldVLASize=1B2026.02 | 0.1 | 27.9 | 41.6 | 43.7 | 17.1 | 10.9 | 38 | 25 | — | — | — | — | — | — | |
| WorldVLAzero-shot=true2026.03 | 0.1 | 27.9 | 41.6 | 43.7 | 17.1 | 10.9 | 38 | 25 | — | — | — | — | — | — | |
| WorldVLAModel Category=World-Action Models, Evaluation Mode=zero-shot, Training Data=standard LIBERO only2026.05 | 0.1 | 27.9 | 41.6 | 43.7 | 17.1 | 10.9 | 38 | 25 | 22 | — | — | — | — | — | |
| WorldVLAtraining=standard LIBERO, protocol=zero-shot2026.07 | 0.1 | 27.9 | 41.6 | 43.7 | 17.1 | 10.9 | 38 | 25 | — | — | — | — | — | — |