ComfyUI v0.42 adds native MOSS-Image support and a Krea 3 node
ComfyUI v0.42.0 adds native core support for the open MOSS-Image text-to-image model and a Krea 3 Image partner node. A v0.42.1 fix followed a few hours later for crashes some users hit with the MOSS-Image template workflow.
What it does
MOSS-Image support, added by Kijai, lets the open model run in a local graph with ComfyUI's own nodes, and the bundled workflow templates moved to v0.11.89. Krea 3 Image arrives as a Partner Node, which runs on Comfy credits rather than on your GPU.
The release also speeds up Qwen 3.x models, fixes Qwen think-block templates in TextGenerate and the webcam node, refactors 3D model remeshing and adds ARM Linux to the tests.
Results and limits
The release notes list pull requests only and report no measured results. v0.42.1, published on 19 October at 01:33 UTC, fixes the crashes with the MOSS-Image template workflow.
Worth knowing
v0.42.0 was published on 18 October at 22:10 UTC (19 October at 00:10 Warsaw time). For faster MOSS-Image renders, community LoRAs cut MOSS-Image-SFT from 50 steps with CFG 3 to 8 steps (lodestones) or 4 steps (darask0's rapid-moss, with its own ComfyUI node); the rapid-moss author measured 4.5 to 6.9 seconds per 1024² image in ComfyUI on an RTX PRO 6000. They have to be applied by that node at runtime, because merging them into bf16 weights wipes out most of the speed-up.
Try it yourself
Update ComfyUI to v0.42.1, open the MOSS-Image template and prompt: "A cinema poster for [film title]: the title '[TITLE]' in large letters at the top, [main character] in the center, small credits text at the bottom."