OmniXtreme: Breaking the Generality Barrier in High-Dynamic Humanoid Control

TL;DR

OmniXtreme introduces a scalable framework that decouples general motor skill learning from physical refinement, enabling high-fidelity, diverse high-dynamic humanoid motion tracking and execution on real robots.

Contribution

It presents a novel decoupled approach combining flow-matching policies with actuation-aware refinement to overcome scalability and physical constraints in humanoid control.

Findings

Maintains high-fidelity tracking across diverse datasets

Successfully executes multiple extreme motions on real robots

Breaks the fidelity-scalability trade-off in high-dynamic control

Abstract

High-fidelity motion tracking serves as the ultimate litmus test for generalizable, human-level motor skills. However, current policies often hit a "generality barrier": as motion libraries scale in diversity, tracking fidelity inevitably collapses - especially for real-world deployment of high-dynamic motions. We identify this failure as the result of two compounding factors: the learning bottleneck in scaling multi-motion optimization and the physical executability constraints that arise in real-world actuation. To overcome these challenges, we introduce OmniXtreme, a scalable framework that decouples general motor skill learning from sim-to-real physical skill refinement. Our approach uses a flow-matching policy with high-capacity architectures to scale representation capacity without interference-intensive multi-motion RL optimization, followed by an actuation-aware refinement phase that ensures robust performance on physical hardware. Extensive experiments demonstrate that OmniXtreme maintains high-fidelity tracking across diverse, high-difficulty datasets. On real robots, the unified policy successfully executes multiple extreme motions, effectively breaking the long-standing fidelity-scalability trade-off in high-dynamic humanoid control.