…icture model. Free for everyone. - New in Images: MiniMax-H3-Image (preview). Type a prompt and get a picture at 4, 8 or 16 megapixels, made on your Mac. - Needs a Mac wi

- Add chainable H3 LoRA nodes with individual filenames and manual strengths for generation and multi-reference editing. - Include two optional LoRA slots in both starter

…kflows using one shared model setup. - Sample one H3 latent frame and include Fizgig's MIT-licensed temporal-group video-VAE decoder with attribution. - Keep one optional

- Added eight-step 768p generation and REF2VA editing workflows with matching Turbo adapters and separate sigma shifts. - Added optional vision-encoder-only ref

… generative detail pass. - Preserved the original H3 image dimensions instead of shrinking output to the model working canvas. - Added a preservation-first Qwen edit prom

MiniMax H3 and workflow updates - Added MiniMax H3 text-to-video, image-to-video, first/last-frame, and native-audio generation through ComfyUI's day-zero native nodes. -