Xiaomi has open-sourced Xiaomi-Robotics-U0, a 38-billion-parameter multimodal autoregressive foundation model that unifies four distinct robot task categories within a single framework1,2. Code and weights are available on GitHub, HuggingFace, and ModelScope.
The release, described as the first unified generative model in embodied AI capable of handling all four task types simultaneously, establishes a complete pipeline for generating and editing robot image and video training data.
Four capabilities in one framework
Xiaomi-Robotics-U0 covers embodied scene generation, embodied transfer, robot interaction video generation, and general-purpose image generation and editing.
Embodied scene generation creates multi-view initial scenes for specified robot hardware from text descriptions, spanning environments from tabletops and kitchens to warehouses and open worlds. Embodied transfer migrates existing robot trajectories to new environments, altering lighting, background, surface materials, target objects, or workspace style while preserving original arm poses and scene layout. Robot interaction video generation produces subsequent video frames from initial observations and operation instructions, maintaining motion coherence and physical consistency with zero-shot generalization to unseen scenarios. General text-to-image and image editing capabilities remain intact, allowing internet-scale visual knowledge to transfer into embodied AI tasks.
Beyond augmenting existing data by swapping objects, lighting, or backgrounds, the model can generate entirely new scenes covering hazardous, extreme, or long-tail environments inaccessible to physical robots.
Benchmark and real-robot results
Xiaomi-Robotics-U0 achieved top scores on the WorldArena benchmark across 126 participating global models. In real robot evaluations under out-of-distribution conditions, strategy task completion rates improved an average of 26% when robots were trained on data augmented by the model.
The model uses a UNIS inference acceleration architecture that improves generation efficiency by approximately 83 times compared to the raw autoregressive paradigm.
Open-source scope
The release includes the full project page, GitHub code repository, HuggingFace model weights, and ModelScope collection. ANALYSIS By open-sourcing both weights and code for a model of this scale, Xiaomi gives external robotics teams a data-generation pipeline that could reduce dependence on costly physical data collection, particularly for edge cases and hazardous environments.
The release follows Xiaomi's robot factory deployment. Xiaomi unveiled the model on Wednesday, according to TechNode.