$Ψ_0$: An Open Foundation Model Towards Universal Humanoid Loco-Manipulation
2026-03-12 • Robotics
Robotics
AI summaryⓘ
The authors created a model called Psi-Zero to help humanoid robots perform complex movement and manipulation tasks better. Instead of training on large amounts of mixed human and robot data, they first trained on clear, first-person human video to learn general actions, then fine-tuned on a smaller set of precise robot movement data. They found this two-step training with high-quality data works better than using much larger but noisier datasets. Their approach improved task success rates significantly with less data. They plan to share all their tools and models with the public.
humanoid robotsloco-manipulationfoundation modelegocentric videoautoregressive pre-trainingflow-based action expertrobot joint controldata efficiencycross-embodiment learning
Authors
Songlin Wei, Hongyi Jing, Boqian Li, Zhenyu Zhao, Jiageng Mao, Zhenhao Ni, Sicheng He, Jie Liu, Xiawei Liu, Kaidi Kang, Sheng Zang, Weiduo Yuan, Marco Pavone, Di Huang, Yue Wang
Abstract
We introduce $Ψ_0$ (Psi-Zero), an open foundation model to address challenging humanoid loco-manipulation tasks. While existing approaches often attempt to address this fundamental problem by co-training on large and diverse human and humanoid data, we argue that this strategy is suboptimal due to the fundamental kinematic and motion disparities between humans and humanoid robots. Therefore, data efficiency and model performance remain unsatisfactory despite the considerable data volume. To address this challenge, \ours\;decouples the learning process to maximize the utility of heterogeneous data sources. Specifically, we propose a staged training paradigm with different learning objectives: First, we autoregressively pre-train a VLM backbone on large-scale egocentric human videos to acquire generalizable visual-action representations. Then, we post-train a flow-based action expert on high-quality humanoid robot data to learn precise robot joint control. Our research further identifies a critical yet often overlooked data recipe: in contrast to approaches that scale with noisy Internet clips or heterogeneous cross-embodiment robot datasets, we demonstrate that pre-training on high-quality egocentric human manipulation data followed by post-training on domain-specific real-world humanoid trajectories yields superior performance. Extensive real-world experiments demonstrate that \ours\ achieves the best performance using only about 800 hours of human video data and 30 hours of real-world robot data, outperforming baselines pre-trained on more than 10$\times$ as much data by over 40\% in overall success rate across multiple tasks. We will open-source the entire ecosystem to the community, including a data processing and training pipeline, a humanoid foundation model, and a real-time action inference engine.