The embodiment gap is the set of differences between the robot that produced a policy's training data and the robot the policy is asked to control. It covers the number and arrangement of joints (degrees of freedom), the action representation (joint positions or end-effector deltas), gripper type, camera count and placement, control rate, and actuator dynamics. A policy has no built-in way to map actions across these differences; it learns mappings only for the embodiments in its data.
Cross-embodiment pretraining tries to narrow the gap through data diversity. The Open X-Embodiment collaboration (2023) assembled "a dataset from 22 different robots collected through a collaboration between 21 institutions" and reported positive transfer for its RT-X models. Later VLAs make the embodiment explicit. LeRobot's X-VLA documentation starts from the problem that "Robot data is fragmented across platforms, sensors, embodiments, and collection protocols", and X-VLA learns soft-prompt embeddings per hardware setup. GR00T N1.7 is described as a cross-embodiment policy. SmolVLA pads state and action vectors to a fixed width of 32 in LeRobot v0.6.1 so that different robots fit one network.
A common misconception is that a pretrained generalist VLA runs zero-shot on any robot. The OpenVLA-OFT paper (Kim et al., 2025) observes that VLAs "struggle with novel robot setups and require fine-tuning to achieve good performance". Deploying a pretrained checkpoint on a robot absent from its pretraining data, such as an SO-101, normally means collecting demonstrations on that robot and fine-tuning. The gap is distinct from sim-to-real transfer, which concerns one robot in two worlds.