← sources

MOSS-Image at Native 4MP

source: reddit (reddit:comfyui)published 2026-10-24 18:59ingested 2026-10-25 02:39

retained text

Testing Krea 2 prompts on MOSS-Image at Native 4MP (\~25s/img, 16\~24GB VRAM on RTX PRO 6000) – First-try raw outputs, no cherry-picking.

Workflow: Official Built-in ComfyUI template (https://docs.comfy.org/tutorials/image/moss-image/moss-image)

All images in this gallery are first-try results with zero cherry-picking, flaws included. It responds to Krea 2 prompts remarkably well—especially on candid flash lighting, skin micro-textures, and stylized proportions. One thing to note during further testing: the model has massive seed variance. Every seed drastically changes the composition, so if you are aiming for a very specific framing rather than blind first-try runs, seed hunting is definitely required.

document fetched 2026-10-25 00:36 · canonical https://reddit.com/r/comfyui/comments/1xegt1q/mossimage_at_native_4mp