Z Image Turbo Pro Workflow V2 BSY 🔥
Added 2025-12-03 17:55:16 +0000 UTCModel links
text_encoders
qwen_3_4b.safetensors
loras
qinglong_detailedeye_z-imageV2(comfy).safetensors
diffusion_models
z_image_turbo_bf16.safetensors
z_image_turbo-Q4_K_S.gguf
https://huggingface.co/jayn7/Z-Image-Turbo-GGUF/blob/main/z_image_turbo-Q4_K_S.gguf
vae
ae.safetensors
https://huggingface.co/Comfy-Org/z_image_turbo/resolve/main/split_files/vae/ae.safetensors
Model Storage Location
```
📂 ComfyUI/
├── 📂 models/
│ ├── 📂 text_encoders/
│ │ └── qwen_3_4b.safetensors
│ ├── 📂 loras/
│ │ └── qinglong_detailedeye_z-imageV2(comfy).safetensors
│ ├── 📂 diffusion_models/
│ │ └── z_image_turbo_bf16.safetensors and z_image_turbo-Q4_K_S.gguf
│ └── 📂 vae/
│ └── ae.safetensors
```
User-Editable Settings
You can modify the items in yellow boxes. Avoid changing anything else, as it may break the workflow.
Settings
CFG Values: Default is 1.
The last pass uses CFG 1.4 — this adds extra detail but increases render time by a few seconds. Adjust if needed.
Load a maximum of 4 LoRAs to maintain quality.
Use the switches to choose between:
Single Image (manual prompting)
Folder Mode (auto prompting)
Want faster speed?
Use a smaller latent image size
Keep the same size for the second pass
Upscale the latent by 1.5x in the third pass
(This saves time but may reduce quality.)
For more help, join the Discord community:
![Ace Avoe [EXKILLO]](https://samukata.com/istorage/36211.jpg)
