小米发布Xiaomi-Robotics-U0：380亿参数多模态自回归模型统一具身合成
小米推出Xiaomi-Robotics-U0，一个380亿参数的多模态自回归模型，用于统一具身合成。该模型将具身生成视为基础图像与视频生成的延伸，联合优化文生图、图像编辑、具身场景生成、具身迁移及具身视频生成。它是首个支持跨多种机器人形态的高质量多视角场景生成模型，并引入结构化可控具身迁移。该模型在单步与序列生成任务上达到SOTA，在具身场景生成与迁移的人类评估中超越GPT-Image-2.0，在World Arena具身视频生成排名第一，并将pi_0.5在真实世界操控任务上的分布外成功率从36.9%提升至63.2%。代码与检查点已开源。
小米这个 38B 的统一具身生成模型，把图像、编辑、场景生成等全塞进一个框架，并在真实机器人操作任务上把 pi_0.5 的成功率从 36.9% 拉到 63.2%，做具身智能的值得认真看。
Authors:Xinghang Li, Jun Guo, Qiwei Li, Long Qian, Hang Lai, Yueze Wang, Hongyu Yan, Jiahang Cao, Xi Chen, Jingen Qu, Jiaxi Song, Nan Sun, Hanye Zhao, Futeng Liu, Wanli Peng, Heyun Wang, Yunhong Wang, Caoyu Xia, Jack Zhao, Diyun Xiang, Hangjun Ye, Heng Qu, Huaping Liu, Jason Li
Abstract: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 this https URL.
| Subjects: | Robotics (cs.RO); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2607.11643 [cs.RO] |
| (or arXiv:2607.11643v1 [cs.RO] for this version) | |
| https://doi.org/10.48550/arXiv.2607.11643 arXiv-issued DOI via DataCite |
Submission history
From: Xinghang Li [view email]
[v1]
Mon, 13 Jul 2026 14:57:58 UTC (36,658 KB)
来源：HuggingFace Daily Papers（社区热门论文） · arxiv.org