Sber unveils GreenWAM world-action model for robot control
Sber
Sber presented GreenWAM on October 21, 2026, a world-action model trained on data from 20 robot families that predicts up to one minute of video from commanded actions. It scored 67.5% on RoboTwin 2.0 versus 58.4% for Physical Intelligence's pi-0.5 and 86.8% on LIBERO-plus without fine-tuning, and raised task success on Sber's humanoid Green to 95.6% from 88.9% for NVIDIA Cosmos 3 Edge.
Why it matters
It fuses scene prediction and action selection in one model trained across many robot types, a step toward general-purpose robot policies rather than per-hardware controllers.
Importance: 3/5
notable research release with cross-robot benchmarks and a preprint
Sources
media
Сбербанк представил модель GreenWAM, которая управляет роботами и предсказывает изменения окружающего мира
(syndicated from Sber press service)