---
license: apache-2.0
tags:
  - image-generation
  - text-to-image
  - z-image
---

# Z-Image-Turbo_clear

- Guide (External site): [English](https://www.ai-image-journey.com/2025/12/z-image-turbo-clear.html) | [Japanese](https://note.com/ai_image_journey/n/nd1323246e146)
- [Civitai](https://civitai.com/models/2197598?modelVersionId=2474394)
- [Z-Image_clear_vae](https://huggingface.co/easygoing0114/Z-Image_clear_vae)

<div style="display: flex; justify-content: center; align-items: center; gap: 2em;">
  <div>
    <img src="./images/Z-Image-Turbo_FP32_clear_vivid_vae_photoreal.png" alt="Z-Image-Turbo_FP32_clear_vivid_vae_anime" width="400" height="400">
  </div>
  <div>
    <img src="./images/Z-Image-Turbo_FP32_clear_vivid_vae_anime.png" alt="Z-Image-Turbo_FP32_clear_vivid_vae_anime" width="400" height="400">
  </div>
</div>

## Compare to [Z-Image-Turbo (Original)](https://huggingface.co/Tongyi-MAI/Z-Image-Turbo)

<div style="display: flex; justify-content: center; align-items: center; gap: 2em;">
  <div>
    <img src="./images/Z-Image-Turbo_clear_blue_dress.png" alt="Z-Image-Turbo_clear_blue_dress" width="400" height="400">
  </div>
  <div>
    <img src="./images/Z-Image-Turbo_original_blue_dress.png" alt="Z-Image-Turbo_original_blue_dress" width="400" height="400">
  </div>
</div>

<p style="text-align:center;">Z-Image-Turbo_clear | Z-Image-Turbo</p>

<div style="text-align: center; margin-left: auto; margin-right: auto; width: 800px; max-width: 90%;">
  <img src="./images/Z-Image-Turbo_clear_blue_dress_close_up.png" alt="Z-Image-Turbo_clear_blue_dress_close_up" width="800" height="400">
</div>

The **Z-Image-Turbo_clear** model is a fine-tuned version of **Z-Image-Turbo** that improves details compared to the original.

For better results, please use it in conjunction with the dedicated VAE, **[Z-Image_clear_vae](https://huggingface.co/easygoing0114/Z-Image_clear_vae)**.

---

## Which Precision Should I Use?

```
+----------+------+----------+----------+--------------------+-----------+
| Format   | Sign | Exponent | Mantissa | Significant Digits | Precision |
+----------+------+----------+----------+--------------------+-----------+
| FP32     | 1bit | 8bit     | 23bit    | 7–8 digits         | ✅        |
| BF16     | 1bit | 8bit     | 7bit     | 2–3 digits         | 👍        |
| FP16     | 1bit | 5bit     | 10bit    | 3–4 digits         | 👍        |
| FP8(e5m2)| 1bit | 5bit     | 2bit     | 1–2 digits         | ⚠️        |
| FP8(e4m3)| 1bit | 4bit     | 3bit     | 1–2 digits         | ⚠️        |
+----------+------+----------+----------+--------------------+-----------+
```
- **Exponent: Stability**
- **Mantissa: Quality**

### Anime Illustration Comparison

<div style="text-align: center; margin-left: auto; margin-right: auto; width: 800px; max-width: 90%;">
  <img src="./images/FP_compare.png" alt="FP_compare" width="800" height="400">
</div>

### Hand Close-up

<div style="text-align: center; margin-left: auto; margin-right: auto; width: 800px; max-width: 90%;">
  <img src="./images/FP_compare_hand_close_up.png" alt="FP_compare_hand_close_up" width="800" height="400">
</div>

- **FP32:** Highest precision, high VRAM usage
- **BF16:** Most popular in Z-Image-Turbo models, good for LoRA compatibility
- **FP16:** Slightly better precision than BF16, less performance degradation on pre-RTX 2000 or non-NVIDIA GPUs
- **FP8:** Lower precision, low VRAM usage, accelerated on RTX 4000 series and later

If you are unsure which precision to use, it is safest to choose the **BF16** format.

---

## Workflow

<div style="text-align: center; margin-left: auto; margin-right: auto; width: 800px; max-width: 100%;">
  <img src="./images/Z-Image-Turbo_clear_workflow_20251226.png" alt="Z-Image-Turbo_clear_workflow_20251226" width="1200" height="600">
</div>

This PNG image contains a ComfyUI workflow.

---

## Recomended Settings

- Sampler:
  - euler
  - res_multistep
- Scheduler: beta
- Steps: 9
- ModelSamplingAuraFlow: shift 7.5 (3-9)

---

## Upscaler

your choice, examples:

- [4x-SwinIR-L_GAN](https://huggingface.co/easygoing0114/AI_upscalers/tree/main)
- [4x_NMKD-Superscale-SP_178000_G](https://openmodeldb.info/models/4x-NMKD-Superscale)

---

## Tips (External Sites)

- ComfyUI VRAM management: [English](https://www.ai-image-journey.com/2025/03/comfyui-setting.html) | [Japanese](https://note.com/ai_image_journey/n/nc46cfff6c3db)
- Color Prompts: [English](https://www.ai-image-journey.com/2025/10/dyes-color-codes.html) | [Japanese](https://note.com/ai_image_journey/n/nbe7eeb3632c3)
- Color Correct custom nodes in ComfyUI: [English](https://www.ai-image-journey.com/2025/09/comfyui-easygoing-nodes.html) | [Japanese](https://note.com/ai_image_journey/n/n5bb33311b866)

---

## License

- [Apache-2.0](./LICENSE)