P-JiT 21K: an open text-to-image model trained for just $21K

pmttyji · reddit · 2026-11-06

Linum-AI released P-JiT 21K on Hugging Face, a text-to-image model trained for only about $21K. It extends the P-JiT architecture with 610M samples at 512px and 50M at 1024px, then post-trains with a variant of FDFO. Weights, GitHub repo, and a technical blog post are all public.

Related event: Two-person lab trains 2B text-to-image model from scratch for $21K, open-sources it(6 posts)→

Original post →

More from Multimodal

Multimodal channel →