Updated: Dec 1, 2025
base model
UPDATE YOUR COMFYUI AND INSTALL MISSING CUSTOM NODES!!!!!
This workflow utilizes a little more VRAM by using the bf16 model of Z-Image and the SeedVR2 Upscaler with the gguf models instead of the full weight. efficient and generates 4K images on my 3070 in about a minute and a half.
If you have enough VRAM, i highly suggest the SeedVR2 7b Q8 GGUF.
The 7b sharp is also decent but can sometimes OVER sharpen the image. the 3b or 7b sharp models work best for ANIME in my opinion.
PHOTOREALISTIC images need the regular 3b or 7b models.
LINKS:::
SeedVR2 GGUF
https://huggingface.co/cmeka/SeedVR2-GGUF/tree/main
Z-Image (bf16)
https://huggingface.co/Comfy-Org/z_image_turbo/tree/main/split_files/diffusion_models
Text Encoder
https://huggingface.co/Comfy-Org/z_image_turbo/tree/main/split_files/text_encoders
Text Encoder GGUF
https://huggingface.co/Qwen/Qwen3-4B-GGUF/tree/main
VAE
https://huggingface.co/Comfy-Org/z_image_turbo/tree/main/split_files/vae