MiniMax H3 Image is now natively supported in ComfyUI

Описание: It's an open-weights image model that handles text-to-image and reference-image editing in a single checkpoint. Connect reference images to edit; leave them off to generate from a prompt.

What stands out:
• Six sampling steps at guidance 1, with nothing to tune and no negative prompt

• Native high-resolution output up to 4096×4096, with wide formats up to 21:9 that are generated at size, not upscaled

• Multi-reference editing: reference images by name in your prompt

• Prompts up to 4,096 tokens

Run it on your own hardware with the open weights, or through Partner Nodes with no download.

MiniMax-H3-Image: Text to Image
https://github.com/Comfy-Org/workflow...

MiniMax-H3-Image: Image Edit
https://github.com/Comfy-Org/workflow...

MiniMax-H3-Image: Image Edit
https://links.comfy.org/4z9epO5

MiniMax-H3-Image: Text to Image
https://links.comfy.org/4iG1nBm