V-JEPA

Model Family

Metaresearch

Video Joint Embedding Predictive Architecture family for learning world models from video without pixel reconstruction.

Models in this family

Model	Family	Released	Modality	Context	Open source	API	Status
V-JEPA 2	V-JEPA	Jun 11, 2025	video	—	—	—	active
V-JEPA	V-JEPA	Feb 15, 2024	video	—	—	—	legacy

No releases recorded for this entry.