Z-Image-Turbo - Image Generator - ComfyUI & One Click Windows Installer

Z-Image-Turbo is a compact 6B-parameter image model that delivers fast, high-quality results with very few sampling steps, making it ideal for local generation on modest GPUs while still matching or beating much larger models in realism and text rendering. This post shares a plug-and-play ComfyUI workflow plus a one-click Windows installer that handles all dependencies, so you can go from download to generating 1024x1024 images in under a minute without any manual setup.
Free workflow on Civitai: civitai.com
 
Included in the Installer Package
It automatically sets up:
ComfyUI and the necessary custom nodes
Sage Attention 2
Flash Attention 2
Triton for Windows
PyTorch: 2.8.0+cu128
 
Preloaded Models
z_image_turbo-Q5_K_M.gguf model — Hugging Face
Qwen3-4B-UD-Q5_K_XL.gguf text encoder — Hugging Face
z_image_turbo_vae.safetensors VAE model — Hugging Face
2xLexicaRRDBNet_Sharp.pth upscale model — Hugging Face
 
Speed
Generate 1024 x 1024 images in less than 1 minute (9 steps) on an RTX 4050 6 GB VRAM GPU using the Q5_K_M GGUF model. The workflow automatically scales with your system, providing faster generation and higher resolution on GPUs like the RTX 4090 or higher-memory setups.
 
System Requirements
Nvidia RTX 30XX, 40XX, or 50XX series GPU (FP16 support required; GTX 10XX/20XX not tested)
CUDA-compatible GPU with a minimum of 4 GB VRAM (more recommended)
Windows OS
At least 15 GB free storage
 
Usage Notes
You can utilize either the Z-Image-Turbo GGUF models or the standard Z-Image-Turbo diffusion models in the workflow.
Enable or disable workflow sections using the levers in the Fast Groups Bypasser node on the left side of the workflow.
Be sure to disable Text to Image nodes if you want to use Image to Image and vice versa.
Enter your detailed descriptive text prompt for the scene you want to generate.
For best results, enhance your prompt with an LLM.