Robotic Manipulation on LIBERO
99.7Spatial Success RateEO1
Evaluation results
| Method | Links | |||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| EO1Architecture type=2D VLAs, Model size=3B, Wrist camera=✓, Pre-trained=✓, Frozen VLM=×, Train tasks=1002026.05 | 99.7 | 99.8 | 99.2 | 94.8 | 98.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| VLA-AdapterModel Type=Single-task Finetuned Model, Merge Method=-, Merge Part=-, Params (B)=0.68 × 42025.11 | 99.6 | 99.6 | 98.2 | 96.4 | 98.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Fast-WAM-JointEmbodied PT.=false2026.03 | 99.6 | 99.4 | 98.2 | 96.8 | 98.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| VSTATraining Range={0.75, 1, 1.25, 1.5}×, Speed=1.0×2026.06 | 99.6 | 97.6 | 97.4 | 92.8 | 96.9 | — | — | — | — | 151 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 1 | 1 | |
| MergeVLA TSVModel Type=Merged Model, Merge Method=TSV [17] + S, Task Masks (S)=true2025.11 | 99.4 | 97.8 | 74.4 | 54.8 | 81.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DualCoT-VLA2026.03 | 99.4 | 99.8 | 97.8 | 98.2 | 98.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Baseline (single-speed)Speed=1×2026.06 | 99.4 | 95.6 | 96 | 95.8 | 96.7 | — | — | — | — | 152 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 1 | 1 | |
| VSTATraining Range={0.75, 1, 1.25, 1.5}×, Speed=0.75×2026.06 | 99.4 | 95.6 | 97.2 | 93.6 | 96.5 | — | — | — | — | 197 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 0.77 | 0.76 | |
| VSTATraining Range={0.75, 1, 1.25, 1.5}×, Speed=1.25×2026.06 | 99.4 | 98 | 96.6 | 94 | 97 | — | — | — | — | 127 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 1.19 | 1.2 | |
| VSTATraining Range={0.75, 1, 1.25, 1.5}×, Speed=1.5×2026.06 | 99.4 | 97.6 | 96.2 | 94 | 96.8 | — | — | — | — | 111 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 1.36 | 1.43 | |
| OmniVLA-RL2026.04 | 99.2 | 99.2 | 98.5 | 93.5 | 97.6 | — | — | — | — | — | — | — | — | — | 1 | — | — | — | — | — | — | — | — | — | — | — | — | |
| BehaviorVLA2026.05 | 99.2 | 99.4 | 98.8 | 94.6 | 98 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| VSTATraining Range={0.5, 1, 1.5, 2}×, Speed=1.0×2026.06 | 99.2 | 95.2 | 97.8 | 94.8 | 96.8 | — | — | — | — | 153 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 1 | 1 | |
| VSTATraining Range={0.5, 1, 1.5, 2}×, Speed=1.5×2026.06 | 99.2 | 98.4 | 96.8 | 94.4 | 97.2 | — | — | — | — | 111 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 1.38 | 1.43 | |
| VSTATraining Range={0.5, 0.75, 1, 1.25, 1.5, 1.75, 2}×, Speed=1.0×2026.06 | 99.2 | 98.2 | 98.4 | 91.8 | 96.9 | — | — | — | — | 153 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 1 | 1 | |
| X-VLA + FASTERFASTER=true2026.03 | 99 | 99.2 | 97.6 | 92 | 97 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| StarVLA-αTraining Protocol=Specialist2026.04 | 99 | 99.8 | 98.5 | 94.1 | 98.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| VGAArchitecture Category=World Action Model Baselines2026.04 | 99 | 99.6 | 98.6 | 95 | 98.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| VLANeXt2026.05 | 99 | 99.2 | 96.6 | 94.6 | 97.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CKT-WAMInitialization=Cosmos-Policy-LIBERO-Predict2-2B2026.05 | 99 | 99.8 | 98.6 | 97.8 | 98.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| GTA-VLA2026.05 | 99 | 98.8 | 98.4 | 97.6 | 98.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| OASISIntermediate=SE(3)-supervised features, Pretrain=false2026.05 | 99 | 98.8 | 97.4 | 95.2 | 97.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| VSTATraining Range={0.5, 0.75, 1, 1.25, 1.5, 1.75, 2}×, Speed=1.25×2026.06 | 99 | 96 | 98.8 | 95.6 | 97.4 | — | — | — | — | 129 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 1.19 | 1.2 | |
| StarVLA-πSteps=30K, Epochs=9.54, VL Backbone=Cosmos-Predict2-2B2026.04 | 98.9 | 98.3 | 94.4 | 90.4 | 95.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| π0.52026.03 | 98.8 | 98.2 | 98 | 92.4 | 96.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| AtomicVLA*Base Model=π0.52026.03 | 98.8 | 98.8 | 97.2 | 96.2 | 97.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| π0.5Embodied PT.=true2026.03 | 98.8 | 98.2 | 98 | 92.4 | 96.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Fast-WAM-IDMEmbodied PT.=false2026.03 | 98.8 | 97.8 | 97.8 | 97.6 | 98 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| π0.5FASTER=false2026.03 | 98.8 | 98.2 | 98 | 92.4 | 96.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| π0.52026.03 | 98.8 | 98.2 | 98 | 92.4 | 96.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| π0.52026.03 | 98.8 | 98.2 | 98 | 92.4 | 96.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MMaDA-VLAWorld Model=true2026.03 | 98.8 | 99.8 | 98 | 95.2 | 98 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| π0.52026.03 | 98.8 | 98.2 | 98 | 92.4 | 96.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| StarVLA-πSteps=30K, Epochs=9.54, VL Backbone=Qwen3-VL-4B2026.04 | 98.8 | 99.6 | 95.8 | 88.4 | 95.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| π0.5Training Protocol=Specialist2026.04 | 98.8 | 98.2 | 98 | 92.4 | 96.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| pi_0.5Architecture Category=Vision-Language-Action Baselines2026.04 | 98.8 | 98.2 | 98 | 92.4 | 96.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| π0.52026.04 | 98.8 | 98.2 | 98 | 92.4 | 96.9 | — | — | — | — | — | — | — | — | — | 2 | — | — | — | — | — | — | — | — | — | — | — | — | |
| π_0.5Evaluation Protocol=Fine-tuned2026.04 | 98.8 | 98.2 | 96.8 | 92.4 | 96.9 | — | — | — | — | — | — | — | — | — | — | 5,200 | 6,300 | 5,800 | 7,800 | 6,300 | — | — | — | — | — | — | — | |
| π0.5Training Protocol=Fine-tuned2026.05 | 98.8 | 98.2 | 98 | 92.4 | 96.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MolmoAct2-ThinkTraining Protocol=Fine-tuned2026.05 | 98.8 | 99.8 | 98.5 | 95.4 | 98.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| pi0.52026.05 | 98.8 | 98.2 | 98 | 92.4 | 96.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ConsisVLA-4D2026.05 | 98.8 | 99.8 | 98 | 95.6 | 98.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| π0.52026.05 | 98.8 | 98.2 | 98 | 92.4 | 96.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| VLA-GSE2026.05 | 98.8 | 99.8 | 98.2 | 96.8 | 98.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| π0.52026.05 | 98.8 | 98.2 | 98 | 92.4 | 96.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| π0.52026.05 | 98.8 | 98.2 | 98 | 92.4 | 96.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| π0.5Architecture type=2D VLAs, Model size=3B, Wrist camera=✓, Pre-trained=✓, Frozen VLM=×, Train tasks=102026.05 | 98.8 | 98.2 | 98 | 92.4 | 96.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| π0.52026.05 | 98.8 | 98.2 | 98 | 92.4 | 96.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| VSTATraining Range={0.5, 1, 1.5, 2}×, Speed=0.5×2026.06 | 98.8 | 88 | 96.4 | 93.3 | 94.1 | — | — | — | — | 295 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 0.52 | 0.5 | |
| π0.52026.06 | 98.8 | 98.2 | 98 | 92.4 | 96.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| StarVLA-αTraining Protocol=Generalist2026.04 | 98.7 | 99.7 | 98.6 | 94.2 | 97.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| VLA-ThinkerArchitecture Category=Vision-Language-Action Baselines2026.04 | 98.7 | 99 | 95.2 | 96.9 | 97.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| π0.5 + FASTERFASTER=true2026.03 | 98.6 | 97.8 | 97.8 | 91.6 | 96.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| StarVLA-OFTSteps=30K, Epochs=9.54, VL Backbone=Cosmos-Predict2-2B2026.04 | 98.6 | 97.6 | 95 | 91.8 | 95.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CogVLA2026.05 | 98.6 | 98.8 | 96.6 | 95.4 | 97.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CogVLA2026.05 | 98.6 | 98.8 | 96.6 | 95.4 | 97.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CogVLA2026.05 | 98.6 | 98.8 | 96.6 | 95.4 | 97.4 | 1 | 2 | 3 | 2 | 2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| VSTATraining Range={0.5, 0.75, 1, 1.25, 1.5, 1.75, 2}×, Speed=1.5×2026.06 | 98.6 | 98 | 96.8 | 95.8 | 97.3 | — | — | — | — | 112 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 1.37 | 1.43 | |
| LingBot-VAEmbodied PT.=true2026.03 | 98.5 | 99.6 | 97.2 | 98.5 | 98.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MemoryVLA2026.03 | 98.4 | 98.4 | 96.4 | 93.4 | 96.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| GeoVLAArchitecture Category=3D Vision-Language-Action Baselines2026.04 | 98.4 | 99 | 96.6 | 96.6 | 97.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ST-πEvaluation Protocol=Fine-tuned2026.04 | 98.4 | 98.3 | 96.9 | 94.3 | 97.3 | — | — | — | — | — | — | — | — | — | — | 5,400 | 6,000 | 5,300 | 6,900 | 5,900 | — | — | — | — | — | — | — | |
| MemoryVLA2026.05 | 98.4 | 98.4 | 96.4 | 93.4 | 96.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| VSTATraining Range={0.5, 0.75, 1, 1.25, 1.5, 1.75, 2}×, Speed=0.75×2026.06 | 98.4 | 95.4 | 97 | 94.4 | 96.3 | — | — | — | — | 201 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 0.76 | 0.76 | |
| NIAFSize=0.77B, Pretraining=false2026.03 | 98.2 | 100 | 98 | 95.5 | 97.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Fast-WAMEmbodied PT.=false2026.03 | 98.2 | 100 | 97 | 95.2 | 97.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| F1Continuous-action=true, World Model=true2026.03 | 98.2 | 97.8 | 95.4 | 91.3 | 95.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| 3D-CAVLAPolicy Setup=Third person camera + Wrist camera + Robot states + Language Instruction2025.05 | 98.2 | 99.8 | 98.2 | 96.1 | 98.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| F12026.04 | 98.2 | 97.8 | 95.4 | 91.3 | 95.7 | — | — | — | — | — | — | — | — | — | 3 | — | — | — | — | — | — | — | — | — | — | — | — | |
| X-VLA2026.05 | 98.2 | 98.6 | 97.8 | 97.6 | 98.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| F1-VLAKV cache=true, Params=4.2B, FLOPs=5.93T, Thrpt. (Hz)=1.682026.05 | 98.2 | 97.8 | 95.4 | 91.3 | 95.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| X-VLAMaximum inference horizon=120 steps2026.05 | 98.2 | 98.6 | 97.8 | 97.6 | 98.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DM0Params=2.4B, Robo-Pretrain=Yes2026.05 | 98.2 | 98.8 | 96.6 | 82.6 | 94.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| X-VLAArchitecture type=2D VLAs, Model size=0.9B, Wrist camera=✓, Pre-trained=✓, Frozen VLM=×, Train tasks=?2026.05 | 98.2 | 98.6 | 97.8 | 97.6 | 98.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Cosmos Policy2026.05 | 98.1 | 100 | 98.2 | 97.6 | 98.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| π0.5 + KIKnowledge Insulation=true, Initialization=from generalist model2025.07 | 98 | 97.8 | 95.6 | 85.8 | 94.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| OpenVLA-OFTViewpoint=Original2025.09 | 98 | 98 | 98 | 91 | 96.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MergeVLAModel Type=Single-task Finetuned Model, Merge Method=-, Merge Part=-, Params (B)=0.68 × 42025.11 | 98 | 98.6 | 95 | 95 | 96.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MergeVLA TAModel Type=Merged Model, Merge Method=TA [21] + S, Merge Part=Except HL→L, Params (B)=0.70, Task Masks (S)=true2025.11 | 98 | 98.8 | 85.4 | 76.6 | 89.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DySL-VLA 7B# Fine-tuned Parameters=226M2026.02 | 98 | 98.2 | 97 | 92.6 | 96.5 | — | — | — | — | — | 27.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| π02026.03 | 98 | 96.8 | 94.4 | 88.4 | 94.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| InternVLA-M12026.04 | 98 | 99 | 93.8 | 92.6 | 95.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| OFlow2026.04 | 98 | 99 | 95 | 94.5 | 96.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| π0.5Params=3.5B, Robo-Pretrain=Yes2026.05 | 98 | 97.8 | 95.6 | 85.8 | 94.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| NORA-1.5Params=3B, Robo-Pretrain=Yes2026.05 | 98 | 96 | 95.4 | 90.5 | 95 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| InternVLA-M12026.05 | 98 | 99 | 93.8 | 92.6 | 95.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| π0.5-KI2026.05 | 98 | 97.8 | 95.6 | 85.8 | 96 | 2 | 6 | 5 | 6 | 4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| X-VLA*FASTER=false, Cluster Evaluated=true2026.03 | 97.8 | 99.4 | 97.8 | 96.8 | 98 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| VLA-Adapter2026.03 | 97.8 | 99.2 | 97.2 | 95 | 97.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| StarVLA-OFTSteps=30K, Epochs=9.54, VL Backbone=Qwen3-VL-4B2026.04 | 97.8 | 98.6 | 96.2 | 93.8 | 96.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| StarVLA-GR00TSteps=30K, Epochs=9.54, VL Backbone=Qwen3-VL-4B2026.04 | 97.8 | 98.8 | 97.4 | 92 | 96.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MolmoAct2Training Protocol=Fine-tuned2026.05 | 97.8 | 100 | 97.8 | 93.2 | 97.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| GR00T-N1.62026.03 | 97.7 | 98.5 | 97.5 | 94.4 | 97 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| GR00T-N1.6Architecture Category=Vision-Language-Action Baselines2026.04 | 97.7 | 98.5 | 97.5 | 94.4 | 97 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| GR00T N1.7Training Protocol=Fine-tuned2026.05 | 97.7 | 97.5 | 98.5 | 94.4 | 97 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| GR00T-N1.62026.05 | 97.7 | 98.5 | 97.5 | 94.4 | 97 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| GR00T-N1.6Architecture type=2D VLAs, Model size=3B, Wrist camera=✓, Pre-trained=✓, Frozen VLM=✓, Train tasks=102026.05 | 97.7 | 98.5 | 97.5 | 94.4 | 97 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| OpenVLA-OFT-7BModel scale=7B, Architecture=OFT2025.07 | 97.6 | 98.4 | 97.9 | 94.5 | 97.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MergeVLA WUDIModel Type=Merged Model, Merge Method=WUDI [8] + S, Task Masks (S)=true2025.11 | 97.6 | 98.2 | 85.6 | 78.2 | 89.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| OpenVLA-OFT 7B# Fine-tuned Parameters=–2026.02 | 97.6 | 98.4 | 97.9 | 94.5 | 97.1 | — | — | — | — | — | 53 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — |