The Beijing Innovation Center for Humanoid Robots has triumphantly introduced China's inaugural open-source Embodied Vision-Language-Action (VLA) large model, XR-1, which adheres to national standards for embodied intelligence. This achievement heralds a major leap forward for China in the realm of intelligent robotics.

XR-1 is underpinned by a robust data infrastructure, RoboMIND2.0, along with a high-fidelity digital asset dataset, ArtVIP, both of which furnish indispensable data support for developers and researchers. Dubbed the 'cerebellum' of embodied intelligence, XR-1 boasts cross-platform operational prowess, empowering it to execute intricate tasks across diverse robot platforms seamlessly.

Moreover, the center has meticulously crafted a comprehensive ecosystem encompassing 'brain + cerebellum + body'. By open-sourcing XR-1, the center has established a novel research platform for both domestic and international researchers and developers, fostering an environment where more groundbreaking applications are anticipated to surface in the foreseeable future.