ComfyUI Wiki · 工作流与开源创作分享·· 2026-10-19AI score58
lodestones releases Kromoss, a fully fine-tuned MOSS-Image model
Text to image
AI briefing
lodestones released Kromoss on October 19, a fully fine-tuned model of MOSS-Image-Base aimed mainly at the Base model's blurriness issue. It uses the Qwen3.5-9B text encoder and FLUX.2 small VAE, and can be dropped directly into existing MOSS-Image workflows.
Open original comfyui-wiki.com
Image GenerationComfyUI & WorkflowsModel Releases#Model release#Image generation#ComfyUI & workflows#Hugging Face