--- format: "aidr-story-markdown/v1" id: "51c446b3095caca33f9eb5a716804e184eb967cbc9f408c0ba1e23aef7271e1d" canonical_url: "https://aidr.today/51c446b3?lang=en" title: "Meta's 8B Parameter JEPA Establishes Robot World Model Scaling Laws" lang: "en" requested_lang: "en" available_langs: ["en","vi"] translation_fallback: null fallback_fields: [] published_at: "2026-10-08T10:39:38.000Z" category: "Research" topics: ["meta","agent"] source_urls: ["https://marketbrief.now/ai/update-metas-8b-parameter-jepa-establishes-robot-world-model-scaling-law-a3fea027","https://x.com/arankomatsuzaki/status/2107470948265779593","https://x.com/BasileTerv987/status/2107551177718284773","https://x.com/randall_balestr/status/2107479655150436504","https://huggingnews.com/ai/update-metas-8b-parameter-jepa-establishes-robot-world-model-scaling-law-a3fea027"] summary: "Meta trained an 8 billion parameter JEPA model that demonstrates robot world models follow scaling laws similar to large language models, and AMI released H-JEPA, a hierarchical system using three stacked JEPAs that improved success on the Visual Antmaze benchmark from 18% to 73%." --- # Meta's 8B Parameter JEPA Establishes Robot World Model Scaling Laws > [Open the canonical story]() **Published:** 2026-10-08T10:39:38.000Z **Category:** Research **Topics:** meta, agent ## Summary Meta trained an 8 billion parameter JEPA model that demonstrates robot world models follow scaling laws similar to large language models, and AMI released H\-JEPA, a hierarchical system using three stacked JEPAs that improved success on the Visual Antmaze benchmark from 18% to 73%\. ## Sources - [Story source]() - [Story source]() - [Story source]() - [Supporting source]() - [Story source]()