Zetta $ζ$: An Efficient Closed-Loop Embodied Harness for Self-Evolving Physical Intelligence
2026-08-17 • Robotics
Robotics
AI summaryⓘ
The authors present Zetta, a system that helps robots learn and adapt while they are working, rather than only reviewing performance afterward. Unlike older methods that follow fixed actions, Zetta updates its behavior in real-time by running multiple loops that handle quick decisions, skill improvements, and validation. When combined with their rollout infrastructure, Z-Infra, Zetta performs very well on robot tasks, improving speed and success rates, and even shows signs of spontaneous problem-solving. This approach suggests a new way for robots to become smarter and more reliable by evolving their skills continuously during physical interactions.
embodied agentsclosed-loop learningruntime criticsrecovery skillspolicy modelsrobot-environment interactionrollout infrastructurezero-shot transferinference speedupself-exploration
Authors
Xin Ding, Liang Mi, Mingzhe Huang, Zixuan Wang, Chao Zhang, Zixu Hao, Fu Chen, Xiangyu Li, Yikai Zheng, Yaoyu Guo, Weijun Wang, Kun Li, Hao Wu, Yunxin Liu, Ting Cao
Abstract
Embodied agents are increasingly used to close the gap left by end-to-end policy models. Yet the agentic path has not realized closed-loop learning in physical execution: existing harnesses remain largely open-loop, following fixed skills during rollout and reflecting only after an episode completes. Such post-hoc reflection cannot govern execution as it unfolds, because physical interaction requires decisions to track rapidly changing robot-environment states at a frequency beyond today's large agentic models. We present Zetta, a closed-loop embodied harness that evolves code-based runtime critics and recovery skills online while keeping the base policy frozen. Through three timescale-separated loops, Zetta provides action-frequency governance, rollout-level critic-recovery proposal, and validation-gated skill updates. Together with Z-Infra, a rollout infrastructure decoupling agent logic from heterogeneous execution resources, Zetta achieves state-of-the-art success on LIBERO-Pro and RoboCasa under our current rollout budget, reaching 90.8% and 93.6%, with an 11.1x inference speedup; success continues to scale with self-exploration experience; learned skills transfer zero-shot, and clear robotic "Aha Moments" emerge. These results show that closed-loop harness self-evolution opens a scaling path for reliable physical intelligence.