Model family
V-JEPA 2
Family overview — summarizes releases; licenses and availability belong to each release.
Maintained by Meta (FAIR)13
V-JEPA 2 is Meta's family of self-supervised video encoders, pretrained on video with a masked latent-feature prediction objective. Meta released it in June 2025 with ViT-L, ViT-H, and ViT-g encoders and an action-conditioned world model (V-JEPA 2-AC) post-trained on robot video, and in March 2026 added V-JEPA 2.1, a retrained family (ViT-B to a 2B-parameter ViT-G) aimed at dense, temporally consistent features.1345
- Repository: vjepa2 repository (external site: github.com)
- Model hub: V-JEPA 2 collection on Hugging Face (external site: huggingface.co)
- Paper: V-JEPA 2 paper (arXiv 2506.09985) (external site: arxiv.org)
- Paper: V-JEPA 2.1 paper (arXiv 2603.14482) (external site: arxiv.org)
Last reviewedEntry updated Documented release
Releases assessed
Each release is assessed on its own. This family has no family-wide openness label.
| Release | Weights | Tier (USASI rubric v0.1) | License | Released |
|---|---|---|---|---|
| V-JEPA 2 ViT-g/16 (384 px) | Weights: Public | Model-disclosure tier (USASI rubric v0.1): Open-stack | MIT, Apache-2.0 | Jun 11, 2025 |
| V-JEPA 2.1 ViT-G/16 (384 px) | Weights: Public | Model-disclosure tier (USASI rubric v0.1): Open-weight | MIT | Mar 16, 2026 |
What it is useful for
Organization context
U.S. eligibility
Project eligibility rests on documented governing or maintaining entities, not on contributors.
Sources
This listing is not an endorsement, a safety assessment, or a federal approval.