@NVIDIAAIDev just dropped Lyra 2.0 — and it turns a single image into a persistent 3D world you can actually explore.

Generating large, consistent 3D environments has been a major weakness for current models: they forget layouts, lose track of motion, and objects wobble or blur… pic.twitter.com/gsq1295QYa

— Futurepedia – Learn to Leverage AI (@futurepedia_io) 16 avril 2026

@NVIDIAAIDev just dropped Lyra 2.0 — and it turns a single image into a persistent 3D world you can actually explore. Generating large, consistent 3D environments has been a major weakness for current models: they forget layouts, lose track of motion, and objects wobble or blur

→ View original post on X — @futurepedia_io