⬢github Python · 1.2K ★ +209 since we first saw it · pushed 5 h ago · Apache-2.0
OpenMOSS/MOSS-Image
Official repository for MOSS-Image, a 5B-parameter base DiT model with high-quality text rendering, layout generation, and efficient fine-tuning.
MOSS-Image is a 5B-parameter text-to-image model from OpenMOSS built on a DiT architecture. It emphasizes accurate text rendering inside images, controllable layout placement, long prompts up to 4K tokens, and strong anime/photorealistic styles. Base and SFT weights are public, with day-0 support in SGLang, vLLM-Omni, and ComfyUI.
Why now: It just launched (October 2026) with open weights, a technical report on arXiv, a Hugging Face demo, and same-day integrations in major inference and workflow tools, quickly earning ~945 stars.
Who it is for: Developers and researchers building text-to-image applications, especially those needing reliable in-image text, layout control, or fine-tuning/LoRA adaptation.
controllable-generationfoundation-modelsimage-generationtext-renderingtext-to-image
Stars over our 83 snapshots: 945 to 1.2K, since 22 h ago.
Where people talked about it
- ⬢github new repos, most starred 10 min ago
API: https://socialmediatrends-api.osmike.com/v1/repos/OpenMOSS/MOSS-Image