---
title: MiniMax-H3_comfy
canonical_url: "https://www.modelscope.cn/models/Kijai/MiniMax-H3_comfy"
md_url: "https://www.modelscope.cn/models/Kijai/MiniMax-H3_comfy.md"
repository: Kijai/MiniMax-H3_comfy
last_updated: 2026-09-13
parameters: 1.8B
tensor_type:
  - BF16
library_name:
  - safetensors
  - pytorch
frameworks:
  - pytorch
downloads: 4285
stars: 13
---

# MiniMax-H3_comfy

> MiniMax-H3_comfy - Kijai 在 ModelScope 开源的模型。Various MiniMax-H3 models converted for ComfyUI usage

- **Repository**: Kijai/MiniMax-H3_comfy
- **Parameters**: 1.8B
- **Downloads**: 4285
- **Stars**: 13
- **Last updated**: 2026-09-13

Source: https://www.modelscope.cn/models/Kijai/MiniMax-H3_comfy

---

## Various MiniMax-H3 models converted for ComfyUI usage

---

# Lightx2v distill lora:

https://huggingface.co/lightx2v/Minimax-h3-Turbo/tree/main

Not 100% sure of the intended alpha, try using this one at lower strength if you get noisy outputs.

`4 steps, 0.75 LoRA strength, er_sde:`

<video controls width=50% src=https://cdn-uploads.huggingface.co/production/uploads/63297908f0b2fc94904a65b8/2k-mKHz4D0-88lC-ELiv9.mp4></video>

`4 steps, 0.75 LoRA strength, sa_solver:`

<video controls width=50% src=https://cdn-uploads.huggingface.co/production/uploads/63297908f0b2fc94904a65b8/kr1DyCMsfPP4EwY9Butlp.mp4></video>

---
