# Z Image Turbo

> Hosted workspace, REST API, and guides for Tongyi-MAI Z-Image-Turbo — a 6B-parameter, 8-step text-to-image model. The site brand is Z Image Turbo (zimageturbo.ai). This is not the official Tongyi Lab site.

## Product

- [Z Image Turbo home](https://zimageturbo.ai/): Browser studio for the 6B / 8-step model. Hosted jobs typically finish in 2–4 seconds.
- [Z-Image-Turbo tool](https://zimageturbo.ai/z-image-turbo): Model-word landing page — speed, bilingual prompts, hosted vs self-host.
- [Z-Image Edit](https://zimageturbo.ai/z-image-edit): Natural-language image editing on the same product.
- [Pricing](https://zimageturbo.ai/pricing): Hosted credit plans. Self-hosted Apache weights stay free.
- [API docs](https://zimageturbo.ai/docs): REST generate/status endpoints and parameters.

## Guides

- [Z Image Turbo prompts](https://zimageturbo.ai/z-image-turbo-prompts): Prompt patterns that stay stable on the 8-step distilled sampler.
- [ComfyUI workflow](https://zimageturbo.ai/z-image-turbo-comfyui-workflow): Local nodes, Shift/CFG/step limits, and troubleshooting.
- [Image editing guide](https://zimageturbo.ai/z-image-turbo-image-editing): Instruction-based edits without leaving the 8–10 step window.

## Comparisons

- [Z Image Turbo vs Nano Banana Lite](https://zimageturbo.ai/z-image-turbo-vs-nano-banana-lite): Lightweight-model trade-offs (speed, bilingual text, self-host).
- [Nano Banana 2 vs GPT Image 2](https://zimageturbo.ai/nanobanana-2-vs-gpt-image-2): Separate model comparison on this site.

## Related tools on this site

- [GPT Image 2](https://zimageturbo.ai/gpt-image-2)
- [Gemini Omni video](https://zimageturbo.ai/gemini-omni-video-generator)
- [Blog](https://zimageturbo.ai/blog)

## Key facts

- Product / site name: Z Image Turbo (also Z-Image-Turbo; older suffix Image Creator AI)
- Domain: https://zimageturbo.ai
- Model: Tongyi-MAI Z-Image-Turbo, 6B S3-DiT, 8 NFEs (ComfyUI may show 9, effective 8)
- Hosted latency (this site): typically 2–4 seconds per image
- Self-host: mid-range GPU usually under 10 seconds; 8GB VRAM possible with GGUF
- Languages: English and Chinese prompts; in-image Chinese typography is a stated strength
- Recommended local settings: CFG 1.0, Shift 3–6, 8–10 steps
- License of official weights: Apache (see Hugging Face)
- Official weights: https://huggingface.co/Tongyi-MAI/Z-Image-Turbo
- Support: support@zimageturbo.ai

## Do not confuse

- zimageturbo.ai is a third-party hosted product and tutorial cluster around the open Z-Image-Turbo weights.
- It is not Tongyi Lab, Alibaba Cloud Model Studio, or Hugging Face.

## Optional

- [Privacy](https://zimageturbo.ai/privacy)
- [Terms](https://zimageturbo.ai/terms)