Free for humans·Paid for agents · x402
Foundation Models for Whole-Body Humanoid Control
arXiv:2504.11002
Robotics Foundation Lab
Free plain-English explainer
Transformer policies trained across morphologies and terrains hint that general-purpose robot brains may arrive the same way general-purpose language models did — by scaling data and architectures together.
Read free explainer →We train a large-scale transformer policy on heterogeneous humanoid datasets that zero-shot adapts to new morphologies and terrains, combining imitation, reinforcement learning, and sim-to-real residual learning.