X-Humanoid Releases TG-VLA Whole-Body VLA Framework

rohanpaul_ai · x · 2026-07-05

X-Humanoid has released TG-VLA, claiming it to be the world's first whole-body Vision-Language-Action (VLA) framework for full-size humanoid robots. Instead of just improving hand control, it enables the entire robot body to operate as a coordinated system, addressing the current difficulties in synchronizing the torso, legs, arms, hands, balance, vision, memory, and timing. The framework includes core components like HEX to achieve cross-scenario learning.

Related event: X-Humanoid Releases TG-VLA: First Full-Body VLA Framework(2 posts)→

Original post →

More from Embodied

Embodied channel →