← Leaderboard

XR-1

VLA
X-Humanoid · Nov 4, 2025

VLA from the Beijing Humanoid Robot Innovation Center with PKU and Beihang that learns unified vision-motion codes as an intermediate representation. Results are self-reported: 72.0% over 20 Tien Kung 2.0 tasks vs 41.0% for π0.5 and 38.0% for GR00T N1.5.

Type
VLA
Parameters
—
Backbone
PaliGemma (π0-style) with unified vision-motion codes
Weights
Open
License
Apache-2.0
Commercial use
Allowed
Released
Nov 4, 2025
Robots
Tien Kung 1.0 / 2.0, UR5e (single and dual), Dual Franka, AgileX
SO-101
–
LeRobot
–
Inputs
–
VRAM
–

Benchmark results

0 of 13 benchmarks

No benchmark results recorded yet.

Not reported on 13 benchmarks
Wrong or missing number? Report an error