README.md

metadata

license: other
license_name: minimax-h3-community-license
license_link: https://huggingface.co/Comfy-Org/MiniMax-H3
base_model:
  - MiniMaxAI/MiniMax-H3
tags:
  - comfyui
  - comfy-native
  - comfy-quant
  - minimax-h3
  - z-image
  - video-generation
  - audio-video
  - merge
pipeline_tag: text-to-video

MiniMax-H3 × Z-Image — spatial detail graft (comfy-native)

The comfy-native cuts of the MiniMax-H3 × Z-Image graft: Z-Image's spatial-attention profile on H3's engine — richer sets and textures, same identity, no per-shot sharpening creep. Full story, demos and verification on the GGUF page.

Load with the plain Load Diffusion Model node, ComfyUI 0.32+. Files are the pruned H3 builds with the graft baked in (zs05 = late-block gains, dose 0.5):

• bf16 — the master (ref2va)
• comfy-fp8 / fp8e5m2 — fp8 scaled
• comfy-int8 / int8_convrot — the fast pick on RTX 50
• comfy-w4a8 / w4a4 / nvfp4 — 4-bit family for 16 GB cards (w4a8 is the quality pick; nvfp4 is Blackwell-native, emulated elsewhere)
• comfy-mxfp8 — 8-bit microscaling, Blackwell-specialized; little benefit on older architectures
• fl2va and ref2va variants where both exist; picking rules are identical to stock H3

RTX 30/40: the GGUF repo is 4–8× faster than any 4-bit comfy-native arm on Ampere.

Made by joeygambino. Questions: open a discussion — I answer.