Physical AI Lab // Episode 3.1 // One-pager

Inside VLAs: three models

Vision encoder, language backbone, action head

Three key points

  1. Same skeleton: all three VLAs look at images, understand language, then output actions directly, with three blocks inside: seeing the world, understanding language, generating actions.
  2. Three action-head designs: OpenVLA splits actions into 256 bins as tokens (R); π-zero uses flow matching with an action expert, and with action chunking runs control at up to 50 hertz (R); GR00T N1's fast system is a diffusion transformer, also trained with flow matching, 120 hertz (R).
  3. Different backbones: OpenVLA = Llama 2 + DINOv2/SigLIP, 7 billion parameters (R); π-zero = PaliGemma, about 3.3 billion (R); GR00T N1 = Eagle-2 (SmolLM2 + SigLIP-2), publicly released GR00T-N1-2B, 2.2 billion (R). Names don't say which is stronger.

Evidence: O official · R research · M media

R256 binsOpenVLA's action head: each action dimension is discretised into 256 bins and predicted as tokens
RFlow matchingπ-zero's action head: on a pre-trained vision-language model, flow matching (a variant of diffusion) generates continuous actions
R50 hertzWith action chunking plus flow matching, π-zero runs control at up to 50 hertz (per the paper; peak, for dexterous tasks like folding laundry)
RFast + slow systemsGR00T N1 as a whole is a dual system: the slow System 2 (vision-language module) understands, and the fast System 1 (diffusion transformer action module) outputs actions in real time
R120 hertzGR00T N1's fast system (a diffusion transformer) outputs closed-loop actions at up to 120 hertz (paper)
RFlow matchingGR00T N1's fast system (a diffusion transformer, DiT) is also trained with flow matching to generate actions (paper)
RLlama 2OpenVLA's language backbone is Llama 2; its vision encoder fuses DINOv2 and SigLIP features
R7 billionparameters in OpenVLA (7B); trained on about 970,000 real robot demonstrations (paper)
RPaliGemmais π-zero's vision-language backbone; its weights are initialised from this open model (paper)
R3.3 billionπ-zero has about 3.3 billion parameters in total (a 3 billion VLM + a 300 million action expert, paper)
REagle-2GR00T N1 uses NVIDIA's Eagle-2 as its vision-language backbone (paper)
RSmolLM2 + SigLIP-2Eagle-2 is fine-tuned from the SmolLM2 language model and the SigLIP-2 image encoder (paper)
R2.2 billionGR00T-N1-2B has about 2.2 billion parameters in total, about 1.34 billion of them in the vision-language backbone (paper)

Further reading