V-JEPA 2

V-JEPA 2 is Meta's new world model, trained on video to understand and predict the physical world. It enables zero-shot robot planning and sets SOTA benchmarks in visual understanding. Model, code, and new benchmarks are now open.

V-JEPA 2 is Meta's advanced world model designed to interpret and anticipate physical interactions through video training. It excels in visual understanding and enables zero-shot robot planning without prior task-specific training. The model sets new state-of-the-art benchmarks and is now openly available, including code and benchmarks, for broader research and application development.