Action chunking means a policy outputs a fixed-length sequence of future actions (a chunk) from one observation, and the controller executes some or all of those actions before calling the policy again. It was introduced to robot imitation learning by Zhao, Kumar, Levine and Finn in "Learning Fine-Grained Bimanual Manipulation with Low-Cost Hardware" (2023), together with the ACT policy and the ALOHA teleoperation rig. The paper takes the term from psychology and argues that predicting the next k steps "reduces the effective horizon of the task by k-fold, mitigating compounding errors".
Two numbers matter and are often confused: how many actions the model predicts, and how many the robot executes before the next call (see action horizon). In LeRobot v0.6.1, ACT defaults to chunk_size=100 and n_action_steps=100; SmolVLA and π0 default to 50 and 50. Executing fewer actions than were predicted makes the robot more reactive at the cost of more model calls.
Chunking is now the norm rather than the exception: Diffusion Policy, π0, π0.5, GR00T N1.7 and OpenVLA-OFT all predict chunks.
A common misconception is that chunking solves latency. It does not. A chunk is executed open-loop, so the robot ignores new observations until the next call, and under synchronous inference it pauses while that call runs. ACT's temporal ensembling, asynchronous inference and real-time chunking each address a different part of this problem. Chunking also makes control rate and model call rate two different numbers.