XPENG XPACE: A Unified Embodied World Model for IRON

Loading video
Loading videoXPENG released XPACE, a unified embodied world model for IRON: one shared video backbone learns both the action and the post-action visual future, trained on 5,000 hours of embodied video spanning human egocentric, human video-action, bridge and teleoperation data. Its world simulator perturbs expert trajectories, simulates deviation and recovery, filters the generated examples and feeds recoveries back into policy training.
Category: world-model
Author: @CyberRobooo
Date: 2026-09-17T00:00:00
Duration: 90.952s





