On November 20, the highly anticipated 2025 Zhiyuan Embodied AI Open Day unfolded in Zhongguancun, Beijing. At present, Zhiyuan has successfully forged partnerships with over 30 entities in the burgeoning field of embodied AI.
In the realm of human - computer interaction, Zhiyuan has made significant strides. It has rolled out the native full - duplex speech large model, RoboBrain - Audio, and the lifelong cognitive memory system, RoboBrain - Memory. Notably, the response latency of RoboBrain - Audio stands at a mere approximately 80 milliseconds. This performance marks a substantial improvement over traditional models, showcasing Zhiyuan's cutting - edge technological prowess in the field.
