VLA models
Octo
United States
First fully open-source generalist robot policy (data, weights and training pipeline); its compact diffusion-head transformer became a standard academic baseline for cross-embodiment manipulation.
OrganizationUC Berkeley / Stanford / CMU / Google DeepMind (academic consortium)
CountryUS
Release2024-05
Params (B)0.1
Open weightsyes
Architecturetransformer policy with diffusion action head; flexible observation and action spaces (Octo-Small 27M, Octo-Base 93M)
LicenseMIT
Modalitiesvision, language, action
Embodimentsmanipulator, cross-embodiment
Training data800k robot trajectories from the Open X-Embodiment dataset
Sources : arXiv (2024-05-20)Octo project (2026-07-09)GitHub (2026-07-09)
← Back to comparator · VLA models