Prompt Guide: Z-Image Turbo
Z-Image Turbo is a state-of-the-art 6B single-stream Diffusion Transformer (S3-DiT) optimized for rapid, high-fidelity local generation. Because of its unique architecture, traditional prompting habits—especially relying on a separate negative_prompt field—do not apply.
This guide walks you through how to structure your descriptions, render bilingual text, and use positive-prompt constraints to achieve pixel-perfect control.
1. Understanding the S3-DiT Brain
Unlike classic diffusion models that process text and images separately, Z-Image Turbo uses a single-stream design. Text tokens and image tokens are merged and processed together in a single sequence. This enables extremely strong instruction adherence, but it comes with a few strict parameters:
- No Classifier-Free Guidance (CFG): Z-Image Turbo runs at
guidance_scale = 0.0or1.0. CFG sliders should be kept at 0/1.- No Traditional Negative Prompts: Because CFG is disabled, the model ignores the negative prompt field completely.
- Few-Step Inference: Optimized specifically for exactly 8 steps. Generating at other step counts will not improve quality and may introduce noise.
Since the traditional negative prompt box is ignored, you must control what is excluded (nudity, text, watermarks, clutter) directly through your positive prompt text.
2. Camera-Style Structured Prompting
Z-Image Turbo responds best to long, detailed, and highly structured prompts (between 80 to 250 words). Rather than typing a chaotic string of keywords, structure your descriptions using this layout:
[Shot Type & Composition] + [Subject Appearance] + [Clothing & Colors] + [Environment/Background] + [Lighting] + [Mood] + [Style/Medium] + [Avoid/Cleanliness Constraints]
High-Fidelity Example:
A medium-shot portrait of an adult woman in her 30s, natural look, medium-length brown hair, wearing a dark professional blazer and shirt, standing in a modern office with a soft blurred background, soft diffused daylight from the side, calm confident expression, realistic photography, 50mm lens, shallow depth of field, plain studio background, no logos, no text, no watermark.
3. "Negative" Prompting inside the Positive Prompt
To remove artifacts, watermarks, or unwanted styles, append explicit exclusion clauses near the end of your prompt. The transformer's strong text comprehension respects these directives:
| Unwanted Element | What to Add at the End of the Prompt |
|---|---|
| Watermarks & Text | no text, no watermark, no branding, no logos |
| Background Clutter | plain background, simple clean backdrop, not cluttered |
| Anatomical Errors | correct human anatomy, natural hands and fingers, no extra limbs |
| Distortion | no motion blur, sharp focus, no grainy noise |
4. Mastering Bilingual Text Rendering
One of Z-Image Turbo's greatest strengths is rendering legible text in both English and Chinese. To do this with maximum accuracy, follow these rules:
- Use Double Quotes: Always wrap the exact letters you want to render in English double quotes (
""). - Keep Languages Separate: Do not mix English and Chinese within the same text element.
- Describe Style and Position: Explicitly dictate where the text should go and how it should look.
- Inhibit Extra Text: Add a cleanup clause at the end to prevent the model from generating random letters.
English Poster Example:
A minimalist movie poster, dark blue background, big English title text "THE GREATEST CITY" centered near the top in bold white sans-serif lettering, small subtitle "A story of hope" below it, clean layout, no extra text, no logos, no watermark.
Chinese Poster Example:
A vintage-style food poster, warm texture background, large red Chinese title text "回忆之味" on the top center in stylized calligraphy, no other characters, no watermark.
5. Prompt Templates
Use these pre-configured templates as starting points in the Prompt Editor:
Professional Corporate Portrait
A close-up headshot of an adult man, friendly confident expression, short dark hair, wearing a simple gray business suit over a white shirt, studio portrait, soft diffused daylight from the front, subtle blurred gray background, realistic photography, 85mm lens, shallow depth of field, fully clothed, modest professional outfit, no logos, no text, no watermark, safe for work.
Flat Vector Illustration
A flat vector illustration of an adult person riding a bicycle through a park, stylized trees and simple buildings in the background, limited pastel color palette, clean modern design, no outlines, minimal shading, modest casual clothing, long-sleeved shirt and trousers, fully clothed, cheerful mood, clear blue sky, no text, no watermark, no logos, simple uncluttered background.
Studio Product Shot
A studio product photo of a sleek coffee mug placed on a white table, soft shadow, subtle reflection, bright neutral background, softbox lighting from the top left, realistic photography, extremely sharp details, no hands, no people, no text, no logos, no watermark, plain white background.