X2 Ultra is the most capable model of the AGIBOT X2 series, a compact humanoid robot about 1.31 m tall. It walks with a humanoid gait, performs demanding dance choreography and moves its head in coordination during voice interaction. It supports multimodal interaction through vision, voice, touch and facial expression, and in conversation it takes gestures, environmental data and context into account. An interactive RGB camera enables face recognition, object detection, gesture recognition and semantic perception of the environment. Only the Ultra model has LiDAR, an RGB-D camera, front stereo and rear monocular vision, with which it moves autonomously, plans its route, localises with 3D SLAM and proactively avoids obstacles without remote control; the base X2 model does not support these functions. When its head is touched, the robot responds by nodding or shaking its head. Additional end effectors (OmniHand, OmniPicker) are compatible only with X2 Ultra and are purchased separately, and X2 Ultra can be customised to order through sales. The battery is swappable, and on one charge the robot runs about 2 hours walking at 0.5 m/s.