This repository (Abiray/MiniMax-H3-Pruned-GGUF) provides pruned and quantized GGUF weights for the MiniMax H3 omni-modal generative model. MiniMax H3 is designed for unified multimodal context processing, capable of generating synchronized high-definition video and 32 kHz stereo audio from text, image, audio, and video inputs.

🌟 Key Highlights

VRAM Efficiency: Pruned architecture compressed down to 8.9 GB – 21.6 GB, bringing MiniMax H3 execution to consumer-tier GPUs.
Synchronized Omni-Modal Output: Simultaneous generation of video (24 FPS) and native stereo audio (32 kHz).
Native ComfyUI Support: Directly compatible with standard ComfyUI-GGUF workflows using the native MiniMax backend.
Dual Pipeline Variants: Full quant suites for both FL2VA (First/Last Frame) and Ref2VA (Omni-Reference) modes.

📂 Repository Weights & Quantization Breakdown

Note: For optimal performance, Q4_K_M is recommended for 16 GB GPUs, while Q5_K_M is recommended for GPUs with 24 GB VRAM.

FL2VA Models (First-and-Last-Frame Mode)

Quantization	File Name	Size	Target Hardware / Use Case
Q3_K_M	`MiniMax-H3-FL2VA-Pruned-Q3_K_M.gguf`	8.9 GB	Low VRAM (~12 GB GPUs)
Q4_K_M	`MiniMax-H3-FL2VA-Pruned-Q4_K_M.gguf`	11.6 GB	Recommended Balance (16 GB GPUs)
Q4_K_S	`MiniMax-H3-FL2VA-Pruned-Q4_K_S.gguf`	11.6 GB	Compact 4-bit K-Quant
Q5_K_M	`MiniMax-H3-FL2VA-Pruned-Q5_K_M.gguf`	14.1 GB	High Quality Balance (24 GB GPUs)
Q5_K_S	`MiniMax-H3-FL2VA-Pruned-Q5_K_S.gguf`	14.1 GB	Compact 5-bit K-Quant
Q6_K	`MiniMax-H3-FL2VA-Pruned-Q6_K.gguf`	16.7 GB	Near-Lossless Precision
Q8_0	`MiniMax-H3-FL2VA-Pruned-Q8_0.gguf`	21.6 GB	Maximum Fidelity / Reference

Ref2VA Models (Omni-Reference Mode)

Quantization	File Name	Size	Target Hardware / Use Case
Q3_K_M	`MiniMax-H3-Ref2VA-Pruned-Q3_K_M.gguf`	8.9 GB	Low VRAM (~12 GB GPUs)
Q4_K_M	`MiniMax-H3-Ref2VA-Pruned-Q4_K_M.gguf`	11.6 GB	Recommended Balance (16 GB GPUs)
Q4_K_S	`MiniMax-H3-Ref2VA-Pruned-Q4_K_S.gguf`	11.6 GB	Compact 4-bit K-Quant
Q5_K_M	`MiniMax-H3-Ref2VA-Pruned-Q5_K_M.gguf`	14.1 GB	High Quality Balance (24 GB GPUs)
Q5_K_S	`MiniMax-H3-Ref2VA-Pruned-Q5_K_S.gguf`	14.1 GB	Compact 5-bit K-Quant
Q6_K	`MiniMax-H3-Ref2VA-Pruned-Q6_K.gguf`	16.7 GB	Near-Lossless Precision
Q8_0	`MiniMax-H3-Ref2VA-Pruned-Q8_0.gguf`	21.6 GB	Maximum Fidelity / Reference

⚙️ Quickstart & ComfyUI Deployment

Requirements

ComfyUI: Version v0.30.0 or higher is required for native MiniMax-H3 architecture support.
Extension: ComfyUI-GGUF custom node package installed.

Setup Steps

Download your desired .gguf variant from the table above.
Place the downloaded .gguf file into the ComfyUI/models/unet/ directory.
In your ComfyUI workflow, load the model using the UnetLoaderGGUF node.

📋 Model Variants & Input Specifications

H3-Base-FL2VA (First-and-Last-Frame Mode):
- No image input: Operates as standard Text-to-Video / Text-to-Audio-Video.
- Single image input: Generates video conditioned on the initial frame or final frame.
- Dual image input: Interpolates seamlessly between designated first and last reference frames.
H3-Base-Ref2VA (Omni-Reference Mode):
- Images: Up to 9 reference images.
- Videos: Up to 3 video clips (2–15 seconds per clip; total combined duration $\le$ 15s).
- Audio: Up to 3 audio clips (must be accompanied by visual input; duration 2–15s).
- Combined Context: Maximum of 12 multi-modal input assets allowed per prompt.

📊 Output Specifications

Duration: 4 – 15 seconds.
Aspect Ratios: Flexible ratio support (16:9, 9:16, 4:3, 3:4, 1:1, 21:9).
Native Resolution: 768p base resolution (scalable to 2K via H3-Regenerate-2K).
Frame Rate: 24 FPS.
Audio Track: 32 kHz stereo audio output.
Multilingual Dialogue: Stable support across 11 primary languages (English, Chinese, French, German, Japanese, Korean, Spanish, Russian, Portuguese, Italian, Arabic).

📜 Legal & License

MiniMax H3 is released under the MiniMax H3 Community License Agreement. Please refer to the MiniMaxAI/MiniMax-H3 repository and the repository's LICENSE file for full terms and commercial usage guidelines.