babakarto/z-image-base-gguf

Z-Image Base GGUF

GGUF quantized version of Tongyi-MAI/Z-Image (Alibaba's 6B parameter diffusion model) for use with ComfyUI-GGUF.

Model Information

Property	Value
Base Model	Tongyi-MAI/Z-Image
Architecture	Lumina2 (DiT-based)
Parameters	~6B
Type	Non-distilled (supports CFG, negative prompts, LoRA)
Recommended Steps	28-50

Available Quantizations

File	Size	VRAM Required	Quality
z_image_base_Q8_0.gguf	6.8 GB	~7-8 GB	Best
z_image_base_BF16.gguf	12.4 GB	~13 GB	Original

Usage with ComfyUI

Requirements

1. ComfyUI
2. ComfyUI-GGUF custom nodes

Installation

1. Install ComfyUI-GGUF:

cd ComfyUI/custom_nodes
git clone https://github.com/city96/ComfyUI-GGUF
pip install --upgrade gguf

2. Download the GGUF file and place it in:

ComfyUI/models/unet/

3. Use the "Unet Loader (GGUF)" node instead of the standard model loader.

Credits

• Original Model: Alibaba Tongyi-MAI Team
• GGUF Tools: city96/ComfyUI-GGUF
• Quantization: babakarto

License

Apache 2.0 (same as original Z-Image model)