Image model Jev redraws MaskGIT from first principles by predicting every pixel in parallel
torchcompiled · x · 2026-09-17
Developer anshuc showcased Jev, an image generation model that creates pictures by predicting every pixel in parallel, rather than the usual autoregressive token-by-token approach.
torchcompiled noted this amounts to reinventing MaskGIT from first principles — the masked image generation paradigm Google introduced in 2022, which generates images quickly via parallel prediction of masked tokens. The observation highlights how some newer image models are returning to parallel decoding.
More from Multimodal
- Mistral OCR's Vik Paruchuri launches a new document-extraction benchmark — VikParuchuri · 2026-09-17
- The 2.5-hour fully AI-generated Odyssey movie is 2.5 hours too long — The Verge AI · 2026-09-17
- Maiden USA: an AI-generated sci-fi comedy short — Living_Operation4319 · 2026-09-17
- Training a GRU planner to guide flow diffusion improves motion continuity in a homebrew video model — pixlpa · 2026-09-17
- Cine Sueños trailer unveils a 10-film AI-generated narrative universe — DreamArchitect_ES · 2026-09-17
- Dev Shows Off AAA-Quality Game Content Made With Spawn in Under a Day — TAbrodi · 2026-09-17