README.md

---
license: other
license_name: minimax-h3-community-license
base_model_relation: quantized
pipeline_tag: text-to-video
library_name: diffusers
tags:
- video-generation
- text-to-video
- image-to-video
- audio-video-generation
- quantized
- 4-bit
- int4
- svdquant
- minimax-h3
- quantfunc
- comfyui
---

<div align="center" style="margin-top: 50px;">
  <a href="https://github.com/QuantFunc/ComfyUI-QuantFunc">
    <img src="https://raw.githubusercontent.com/QuantFunc/ComfyUI-QuantFunc/main/assets/logo.webp" width="300" alt="QuantFunc logo">
  </a>
</div>

<p align="center">
  🌐 <a href="https://www.quantfunc.com/">Website</a> &nbsp;|&nbsp;
  🐙 <a href="https://github.com/QuantFunc/ComfyUI-QuantFunc">GitHub</a> &nbsp;|&nbsp;
  🤗 <a href="https://huggingface.co/QuantFunc">Hugging Face</a> &nbsp;|&nbsp;
  🤖 <a href="https://www.modelscope.cn/profile/QuantFunc">ModelScope</a> &nbsp;|&nbsp;
  🎨 <a href="https://civitai.com/user/quantfunc">Civitai</a> &nbsp;|&nbsp;
  🎮 <a href="https://discord.gg/jCp9TpFWcn">Discord</a>
</p>

# MiniMax-H3-QuantFunc-4bit

**4x compression, quality held.**

QuantFunc INT4 cuts MiniMax H3's core weight precision from 16-bit to 4-bit. Character detail, style fidelity and fast motion stay clear and coherent, while H3's native video+audio generation is fully preserved.

In our internal FL2VA evaluation, QuantFunc INT4 vs the BF16 baseline (same prompt, same seed) measures ~**23.7 dB PSNR**.

## Showcase

All clips below were generated by MiniMax-H3-QuantFunc-4bit — click a player to watch.

<table>
  <tr>
    <th>Live-action performance</th>
    <th>Animated chase</th>
  </tr>
  <tr>
    <td>
      <video controls preload="metadata" poster="https://huggingface.co/QuantFunc/Minimax-H3-Quantfunc-4bit/resolve/main/assets/h3_case1.webp" width="100%">
        <source src="https://huggingface.co/QuantFunc/Minimax-H3-Quantfunc-4bit/resolve/main/assets/h3_case1.mp4" type="video/mp4">
      </video>
    </td>
    <td>
      <video controls preload="metadata" poster="https://huggingface.co/QuantFunc/Minimax-H3-Quantfunc-4bit/resolve/main/assets/h3_case2.webp" width="100%">
        <source src="https://huggingface.co/QuantFunc/Minimax-H3-Quantfunc-4bit/resolve/main/assets/h3_case2.mp4" type="video/mp4">
      </video>
    </td>
  </tr>
  <tr>
    <th>High-speed car chase</th>
    <th>Stylized character</th>
  </tr>
  <tr>
    <td>
      <video controls preload="metadata" poster="https://huggingface.co/QuantFunc/Minimax-H3-Quantfunc-4bit/resolve/main/assets/h3_case3.webp" width="100%">
        <source src="https://huggingface.co/QuantFunc/Minimax-H3-Quantfunc-4bit/resolve/main/assets/h3_case3.mp4" type="video/mp4">
      </video>
    </td>
    <td>
      <video controls preload="metadata" poster="https://huggingface.co/QuantFunc/Minimax-H3-Quantfunc-4bit/resolve/main/assets/h3_case4.webp" width="100%">
        <source src="https://huggingface.co/QuantFunc/Minimax-H3-Quantfunc-4bit/resolve/main/assets/h3_case4.mp4" type="video/mp4">
      </video>
    </td>
  </tr>
</table>

> Poster frames are the first frame of each clip. Showcase spec: 896 × 1184, 124 frames, 24 FPS, ~5s.

## Same 124 frames, up to 3.19x FP8's per-step speed

On an **RTX 4090**, 768 × 768, 5s, 124 frames:

| Backend | Per-step time | Relative speed |
|---|---:|---:|
| **QuantFunc INT4** | **3.2s** | — |
| INT8 ConvRot | 8.5s | **2.66x faster** |
| FP8 | 10.2s | **3.19x faster** |

> Per-step core-model inference time only — excludes text encoder, VAE, audio processing and video saving. Actual speed varies with resolution, frame count, driver and software version.

## Swap one loader, keep the rest of your workflow

1. Install or update [ComfyUI-QuantFunc](https://github.com/QuantFunc/ComfyUI-QuantFunc).
2. Download the FL2VA 4-step or Ref2VA 8-step weights.
3. Swap your model loader for the QuantFunc loader and pick the matching weight file.

Prompts, reference images/videos, audio assets and the rest of your workflow nodes stay unchanged.

Both weight sets already have the acceleration LoRA, Token Refiner, INT4 Refiner and INT8 Conv Sidecar fused in — no extra components to attach.

## RTX 20-series through GB300, one build covers it all

Runs on every NVIDIA **SM75+** GPU: RTX 20/30/40/50-series, A100, H100, H200, B100, B200, GB300.

4-bit weights significantly cut the weight-bandwidth cost of loading and inference, making MiniMax H3 much easier to run on consumer GPUs. Actual VRAM needs depend on resolution, frame count, reference-asset count and the rest of your workflow.

## Choose a model

| Variant | File | Size | Recommended steps | Use case |
|---|---|---:|---:|---|
| **FL2VA** | `minimax_h3_fl2va_4step_quantfunc_int4_r128.safetensors` | 12.37 GB | **4 steps** | Text-to-audio/video, first-frame, last-frame, first+last-frame control |
| **Ref2VA** | `minimax_h3_ref2va_8steps_quantfunc_int4_r128.safetensors` | 12.37 GB | **8 steps** | Multimodal reference generation from images, video and audio |

FL2VA supports zero, one or two input images; Ref2VA targets more complex multimodal reference scenarios. See the [official MiniMax H3 repo](https://github.com/MiniMax-AI/MiniMax-H3) for details on both modes.

Use the matching scheduler for the model. Both weight sets already carry the acceleration LoRA fused in — no need to load it separately.

## Loading

Weights use QuantFunc's own sealed safetensors format (quantization parameters and metadata are sealed).

Load with **ComfyUI-QuantFunc** or the **QuantFunc inference engine** — this is not a drop-in Diffusers checkpoint. (This repo declares `library_name: diffusers` purely so Hugging Face counts real downloads of the flat `.safetensors` files above — see [HF's download-stats docs](https://huggingface.co/docs/hub/models-download-stats) — not because the files load via the `diffusers` package.)

If the model isn't recognized or fails to load, update ComfyUI-QuantFunc to the latest version and restart ComfyUI.

## Technical details

- INT4 weights × INT4 activations (W4A4)
- SVDQuant rank 128, group size 64
- INT4 Token Refiner + INT4 Refiner
- INT8 Conv Sidecar
- FL2VA measured max relative output error from fp16 adaLN folding: ~**7.58e-4**
- Minimum GPU architecture: NVIDIA SM75

## Source & license

This repository distributes derived quantized weights produced from [MiniMax H3](https://github.com/MiniMax-AI/MiniMax-H3).

MiniMax H3 and its derivative weights follow the **MiniMax H3 Community License Agreement**. QuantFunc's quantization tooling and implementation follow their own respective licenses. Please read and comply with the original model's license terms before use.

## Community

- 🌐 [QuantFunc website](https://www.quantfunc.com/)
- 🐙 [ComfyUI-QuantFunc on GitHub](https://github.com/QuantFunc/ComfyUI-QuantFunc)
- 🎮 [Discord](https://discord.gg/jCp9TpFWcn)