ω-0 Model Enables Humanoid Robots to Perform Household Tasks with 81.8% Success
CyberRobooo · x · 2026-08-09
Researchers from NTU, PKU, BAAI, and HKUST(GZ) released ω-0 (OMEGA-0), a latent predictive world-action model for humanoid robots. Unlike traditional systems that separate locomotion and manipulation, ω-0 directly generates coordinated whole-body actions.
Trained on the 40+ hour ω-HOME multimodal dataset, the model was deployed on a Unitree G1 robot. Across 11 household tasks such as picking items, wiping tables, and grabbing drinks, it achieved an 81.8% success rate and 90.3% task progress using a single policy.
More from Embodied
- Unitree Unveils G1 Robot Army Starting at $16K for Mass Production — Scobleizer · 2026-08-09
- Faraday Future Demos EAI Patrol System for Autonomous Site Security — Scobleizer · 2026-08-09
- New Research: Optimizing Contact Timing for Fluid Humanoid Running and Gymnastics — Scobleizer · 2026-08-09
- NVIDIA Open-Sources MotionBricks: Real-Time Generation of 350K Character Motions — smtabatabaie · 2026-08-09
- Leg-KILO: Open-Sourcing Robust SLAM for Dynamic Legged Robots — rsasaki0109 · 2026-08-09
- UIUC & CMU's LUCID: Teaching Robots Manipulation Skills from Human Videos — lukas_m_ziegler · 2026-08-09