arXiv Robotics — research abstracts· Gennady Sidorov, Valerii Kornilov, Denis Grankin, Polina Fedotova, Aleksey Postnikov·2026-10-14 04:00 UTC· 2 days agoSignalEditorial score85

GreenWAM: A World Foundation Model for Robotics

Summary

This paper introduces GreenWAM, a world foundation model for robotics that serves both world modeling and policy generation roles. It achieves top-tier performance on WorldArena 1.0 and ranks among the top entries in RoboTwin 2.0. The model is adapted for real-world humanoid robots and distilled into a single-step policy for real-time control, with controlled experiments highlighting the benefits of robot specialization and maintaining world-model modes during policy adaptation.

Editorial context

GreenWAM is presented as a world foundation model for robotics that excels in both world modeling and policy generation, outperforming existing models in several benchmarks. The paper introduces a shared anatomical action space and joint training across robot families, enabling cross-embodiment performance and real-time adaptation.

Source: arXiv Robotics — research abstracts · Read original article ↗

Loading article text…

Source：arXiv Robotics — research abstracts · arxiv.org

#Model release #World models #Sim-to-real #Robot policies