Blog
Making video with generative AI: guides
Plain, practical guides to making video with generative AI: how today's video models work, the files and settings that matter, and the hardware they need.
- Start hereMiniMax H3 in ComfyUI: what every part doesA beginner's map of ComfyUI's MiniMax H3 template, with pictures: the video model, the text encoder, the sampler and the two decoders.
- Guide 2H3's modes explained: T2VA, I2VA, FL2VA, L2VA and Ref2VAA beginner's guide to MiniMax H3's five modes, with pictures: what each takes in and gives back, where it is in ComfyUI, and when to use it.
- Guide 3H3 model file names decoded: bf16, fp8, int8, convrot, nvfp4A beginner's guide to MiniMax H3's file names, with pictures: what pruned, bf16, int8_convrot, fp8_scaled, w6a8 and nvfp4 mean, and which suit your card.
- Guide 4The MiniMax H3 settings that actually matterA beginner's guide to MiniMax H3's settings in ComfyUI, with pictures: size, length, steps, sampler, shift, CFG, speed-up LoRAs and seeds.
- Guide 5Running MiniMax H3 on 16, 24 and 32 GB graphics cardsA beginner's guide, with a chart, to the video memory and RAM MiniMax H3 needs in ComfyUI: offloading, render time, and what to choose for each card.
- Guide 6Your first MiniMax H3 video in ComfyUI: a checklistA step-by-step checklist, with a picture of each click, for a first MiniMax H3 clip in ComfyUI: files and folders, the template, settings, a prompt and fixes.