# MiniMax H3 INT8 R2V Community

## Summary

Community `MiniMax H3` reference-to-video workflow built around `INT8` inference and an explicit acceleration stack.

## Model Family

- `minimax`

## Status

- `imported-only`

## Source

- Workflow repo: <https://huggingface.co/javawock7618/comfy-MiniMax-H3-workflows>
- Workflow file: <https://huggingface.co/javawock7618/comfy-MiniMax-H3-workflows/resolve/main/MiniMax_int8_R2V-javano2608.5.json?download=true>
- Original publisher: `javawock7618`
- Date imported: `2026-08-08`

## Developer notes

- The repo describes this as part of an `INT8 FLF & Reference Video` collection.
- The author explicitly targets:
  - reduced VRAM usage
  - faster inference
  - improved memory efficiency
- The documented acceleration stack includes:
  - `Sol-Attn`
  - `Spectrum`
  - optional `Lightx2v LoRA`
  - optional `MiniMax-H3-Turbo-Lora`
  - optional latent upscaler

## Our notes

- This is the strongest low-VRAM-oriented MiniMax import we have so far because it is built around `INT8` from the start.
- Still not guaranteed easy on a `3070`, but much more aligned with your constraints than generic heavyweight MiniMax examples.
- Good candidate once you want to test whether MiniMax reference-driven video is remotely viable locally.

## Required custom nodes

- `ComfyUI-SolAttn_triton`
- `ComfyUI-Spectrum-MiniMax-H3`
- optional `ComfyUI-H3-Motion-Context` for related FLF workflows
- optional `ComfyUI-MiniMaxH3_LatentUpscaler`
- `Kijai/MiniMax-H3_comfy` for Lightx2v acceleration
- optional `MiniMax-H3-Turbo-Lora-ComfyUI`

## Notes

- Best current MiniMax low-VRAM / INT8 reference-video workflow in the repo.
- More of a carefully chosen experimental path than a guaranteed easy win.