OpenMOSS released MOSS-Image, and it’s now supported in ComfyUI.
图像生成模型支持ComfyUI，非视频模型
用ComfyUI下载MOSS-Image试文字渲染
What it is: 5B DiT, using a Qwen3.5-9B text encoder and the FLUX.2 VAE Two versions: Base , and SFT , which is fine-tuned on curated data Apache 2.0 Why it’s worth trying: Text rendering. Legible text in English and Chinese, which most open models still struggle with. Lay
判断口径 agnes-3.0-flash · 入库 #1072 · 回首页