Real-Time Execution of Action Chunking Flow Policies
This is experimental HTML to improve accessibility. We invite you to report rendering errors. Use Alt+Y to toggle on accessible reporting links and Alt+Shift+Y to toggle off. Learn more about this project and help improve conversions. Modern AI systems, especially those interacting with the physical world, increasingly require real-time performance. However, the high latency of state-of-the-art generalist models, including recent vision-language-action models (VLAs), poses a significant challenge. While action chunking has enabled temporal consistency in high-frequency control tasks, it does not fully address the latency problem, leading to pauses or out-of-distribution jerky movements at chunk boundaries. This paper presents a novel inference-time algorithm that enables smooth asynchronous execution of action chunking policies. Our method, real-time chunking (RTC), is applicable to any diffusion- or flow-based VLA out of the box with no re-training. It generates the next action chunk
Real-Time Execution of Action Chunking Flow Policies Kevin Black 1,2, Manuel Y. Galliker 1 Sergey Levine 1,2 1 Physical Intelligence 2 UC Berkeley {kevin,manuel,sergey}@physicalintelligence.company Abstract Modern AI systems, especially those interacting with the physical world, increasingly require real-time performance. However, the high latency of state-of-the-art generalist models, including recent vision-language-action models (VLAs), poses a significant challenge. While action chunking has enabled temporal consistency in high-frequency control tasks, it does not fully address the latency
Explore this link on the map →related reading
- Real-Time Action Chunking with Large Modelspi.website
- Real-Time Action Chunking with Large Modelsphysicalintelligence.company
- FASTERinnovator-zero.github.io
- Interaction Models: A Scalable Approach to Human-AI Collaboration - Thinking Machines Labthinkingmachines.ai
- Reinforcement Learning with Action Chunkingarxiv.org
- Visualizing Flow Matching in Roboticsabhinavjha.xyz
- Precise Manipulation with Efficient Online RLpi.website
- Composer2.pdfcursor.com
- Improving Composer through real-time RL · Cursorcursor.com
- [2507.07969] Reinforcement Learning with Action Chunkingar5iv.labs.arxiv.org
- Is Frontier Asynchronous RL Solved? — Luke J. Huangluk-huang.github.io
- State of Vision-Language-Action (VLA) Research at ICLR 2026 – Moritz Reussmbreuss.github.io