Flow - 4x faster than typing
Mini toolsXiaomi Robotics U0
By Xiaomi
Xiaomi Robotics U0 unifies foundation image and video generation with embodied world modeling in one autoregressive framework, built from a pretrained world foundation model. It jointly learns text to image generation, image editing, multi view embodied scene generation, controllable embodied transfer, and embodied video generation. A structured control formulation separates workspace, background, foreground objects, target objects, and lighting into independent dimensions, enabling scalable synthetic robot training data. Given a single observation and an instruction, it generates temporally coherent multi view video rollouts. Generated data has been shown to improve real world manipulation success rates when used to augment robot policy training, and the model ranks first on the WorldArena embodied video generation benchmark.
New
Multimodal
Overview
A 38 billion parameter multimodal autoregressive model for unified embodied synthesis. It performs text to image generation, image editing, multi view robot scene generation, embodied scene transfer, and embodied video generation, extending foundation image and video generation into robotics settings.
About Xiaomi
Consumer electronics and smart device company making smartphones, wearables, IoT products, home appliances, smart TVs, scooters, and connected lifestyle hardware.
View Company Profile
Last updated: July 14, 2026
