Stable Diffusion Art 教程· Andrew·· 16 hours agoPicksAI score74
MiniMax H3 Image generates native 16MP images in ComfyUI (with workflow)
MiniMax H3 Image in ComfyUI: Native 16MP Images (Workflow Included)
AI briefing
MiniMax H3 Image can natively generate approximately 16.8MP images in ComfyUI, completed in 6 steps, without external super-resolution. The tutorial provides a complete workflow JSON and gives quantized model options: diffusion model 21GB, text encoder 15.7GB, image VAE 1.17GB, and can run on a 24GB GPU.
Why it matters
The tutorial provides a complete ComfyUI workflow and quantized model options, allowing local reproduction of native 16.8MP image generation, eliminating the external super-resolution step.
Open original stable-diffusion-art.com