# MiniMax H3 INT8 R2V Javawock Pack

## Summary

Same-day `MiniMax H3` reference-to-video workflow from the `javawock7618` pack. This is the most direct new character-consistency path in the bundle.

## Model Family

- `minimax`

## Status

- `imported-only`

## Source

- Workflow pack page: <https://huggingface.co/javawock7618/comfy-MiniMax-H3-workflows>
- Workflow file: <https://huggingface.co/javawock7618/comfy-MiniMax-H3-workflows/raw/main/MiniMax_int8_R2V-javano2608.23.1.json>
- Original publisher: `javawock7618`
- Date imported: `2026-08-22`

## Developer notes

- The pack README describes this as the `Reference-to-Video` workflow.
- It is built around INT8 inference, Sol-Attn, Spectrum, Lightx2v LoRA support, optional Turbo-LoRA, and latent upscaling.
- The workflow uses the pack's dedicated reference LoRA path for appearance retention.

## Our notes

- Best new MiniMax H3 import today if you care about reference retention more than flashy motion.
- Still heavy on an `RTX 3070`, but the INT8 path is the main reason it is worth trying locally.
- Good comparison point against current LTX reference-control workflows.

## Required custom nodes

- `ComfyUI-SolAttn_triton`
- `ComfyUI-Spectrum-MiniMax-H3`
- `ComfyUI-Switchboard`
- `ComfyUI-MiniMaxH3_LatentUpscaler`

## Required models

- `Comfy-Org/MiniMax-H3`
- `minimax_h3_ref_lora_rank_256_bf16.safetensors`
- `lightx2v/Minimax-h3-Turbo` if the accelerated path is enabled
- `drbaph/MiniMax-H3-Turbo-Lora-ComfyUI` if the turbo path is enabled

## Notes

- Best same-day MiniMax workflow for reference-guided video.