HY-Embodied-0.5

HY-Embodied-0.5 is an open-source embodied Vision-Language-Action model from Tencent Hunyuan, part of the HY-Embodied family of embodied foundation models. Trained on 10,000+ hours of real-world robot demonstrations, of which 2,000+ hours have been released. Uses MoT (Mixture-of-Transformers) architecture. Released under Apache-2.0 license. Now part of the broader HY-Embodied open-source ecosystem including Hy-Embodied-VLM-1.0 and Hy-Embodied-RxBrain-1.0.