Xiaomi-Robotics-U0: Unified Embodied Synthesis with World Foundation Model

Hugging Face Daily Papers Papers
embodied-ai world-model robotics multimodal autoregressive-model scene-generation foundation-model

Summary

Xiaomi Robotics introduces U0, a 38-billion-parameter multimodal autoregressive model for unified embodied synthesis, treating embodied generation as an extension of image and video generation. It achieves state-of-the-art results on multiple embodied tasks, outperforming GPT-Image-2.0 and improving real-world manipulation success rates.

Recent foundation image and video generation models offer strong generalization and controllability, but their direct application to embodied scenarios is limited by requirements for multi-view consistency, geometric coherence, and robot embodiment constraints. Existing methods typically adapt foundation models with limited robot data, often sacrificing visual knowledge acquired during large-scale pre-training. We present Xiaomi-Robotics-U0, a 38-billion-parameter multimodal autoregressive model for unified embodied synthesis. It treats embodied generation as an extension of foundation image and video generation and jointly optimizes text-to-image generation, image editing, embodied scene generation, embodied transfer, and embodied video generation. This unified framework preserves the generalization of the pre-trained world foundation model while adapting it to embodied settings. Xiaomi-Robotics-U0 is the first model to support high-quality multi-view scene generation across multiple robot embodiments and to introduce structured, controllable embodied transfer for fine-grained editing while preserving multi-view consistency and interaction dynamics. It achieves state-of-the-art results on single-step and sequential generation tasks, outperforming GPT-Image-2.0 in human evaluations of embodied scene generation and transfer, ranking first on World Arena for embodied video generation, and improving the out-of-distribution success rate of pi_0.5 from 36.9% to 63.2% on challenging real-world manipulation tasks. These results show that foundation world models can serve both as embodied world models and scalable data engines for embodied intelligence. Code and checkpoints are available at https://robotics.xiaomi.com/xiaomi-robotics-u0.html.
Original Article
View Cached Full Text

Cached at: 07/14/26, 04:16 PM

Paper page - Xiaomi-Robotics-U0: Unified Embodied Synthesis with World Foundation Model

Source: https://huggingface.co/papers/2607.11643 Authors:

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

Abstract

Recentfoundationimageandvideogenerationmodelsofferstronggeneralizationandcontrollability,buttheirdirectapplicationtoembodiedscenariosislimitedbyrequirementsformulti-viewconsistency,geometriccoherence,androbotembodimentconstraints.Existingmethodstypicallyadaptfoundationmodelswithlimitedrobotdata,oftensacrificingvisualknowledgeacquiredduringlarge-scalepre-training.WepresentXiaomi-Robotics-U0,a38-billion-parametermultimodalautoregressivemodelforunifiedembodiedsynthesis.Ittreatsembodiedgenerationasanextensionoffoundationimageandvideogenerationandjointlyoptimizestext-to-imagegeneration,imageediting,embodiedscenegeneration,embodiedtransfer,andembodiedvideogeneration.Thisunifiedframeworkpreservesthegeneralizationofthepre-trainedworldfoundationmodelwhileadaptingittoembodiedsettings.Xiaomi-Robotics-U0isthefirstmodeltosupporthigh-qualitymulti-viewscenegenerationacrossmultiplerobotembodimentsandtointroducestructured,controllableembodiedtransferforfine-grainededitingwhilepreservingmulti-viewconsistencyandinteractiondynamics.Itachievesstate-of-the-artresultsonsingle-stepandsequentialgenerationtasks,outperformingGPT-Image-2.0inhumanevaluationsofembodiedscenegenerationandtransfer,rankingfirstonWorldArenaforembodiedvideogeneration,andimprovingtheout-of-distributionsuccessrateofpi_0.5from36.9%to63.2%onchallengingreal-worldmanipulationtasks.Theseresultsshowthatfoundationworldmodelscanservebothasembodiedworldmodelsandscalabledataenginesforembodiedintelligence.Codeandcheckpointsareavailableathttps://robotics.xiaomi.com/xiaomi-robotics-u0.html.

View arXiv pageView PDFAdd to collection

Models citing this paper2

#### XiaomiRobotics/Xiaomi-Robotics-U0 Robotics• 34B• Updatedabout 9 hours ago • 13 • 3 #### XiaomiRobotics/Xiaomi-Robotics-U0-FlashAR Robotics• 38B• Updatedabout 9 hours ago • 5 • 3

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2607.11643 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2607.11643 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

Xiaomi Open-Sources Embodied AI Foundation Model Xiaomi-Robotics-1 (4 minute read)

TLDR AI

Xiaomi open-sourced Xiaomi-Robotics-1, an embodied AI foundation model pretrained on over 100,000 hours of UMI data and post-trained on 10,000+ hours of cross-embodiment data. The release includes the full real-robot post-training and deployment pipeline, aiming to challenge proprietary robotics models from Figure AI and Tesla.

Xiaomi-Robotics-1

Hacker News Top

Xiaomi presents Robotics-1, a robot policy model trained via embodiment-free pre-training on 100,000 hours of data, showing clean scaling behavior and strong generalization to real-world tasks.

Qwen-RobotWorld Technical Report: Unifying Embodied World Modeling through Language-Conditioned Video Generation

Hugging Face Daily Papers

Qwen-RobotWorld is a language-conditioned video world model that predicts future visual trajectories across multiple robotic domains using a double-stream diffusion transformer and an 8.6M video-text corpus. It unifies embodied world modeling for robotic manipulation, autonomous driving, indoor navigation, and human-to-robot transfer, achieving top benchmarks on EWMBench and DreamGen Bench.