Images · Alibaba (Tongyi Lab)

Z-Image Turbo

A small, fast open model with good photorealism, private and about 0.5 credit. Pick it for quick drafts that still look real.

Private Low cost
Latest versionZ-Image-TurboNovember 2025
Artificial Analysis Text to Image Arena941 Elo#35 of 37 on Artificial Analysis Text to Image Arena
Price$0.0045 per imageZ-Image Turbo
Key specEnglish and Chinese text

What sets it apart

A 6B open model from Alibaba's Tongyi Lab built to draw in 8 steps, with sub-second latency on enterprise GPUs. It combines photorealism with English and Chinese text and runs in private mode at about half a credit.

  • Draws in 8 steps, very fast
  • 6B, Apache 2.0, fits 16 GB VRAM
  • English and Chinese text

Recommended for: try many ideas quickly for almost nothing.

Strengths and limits

Where it shines

  • Very fast: built to draw in 8 steps
  • Photorealism and bilingual (English, Chinese) text

Think twice when

  • A 6B model: less detail than the large paid flagships
  • Apache 2.0 license
  • Prompts up to 7,500 characters

Use cases

A fast photoreal draft

A photo of a ceramic teapot and two cups on a wooden tray by a window, soft morning light.

It is built for quick drafts that still look real.

A bilingual sign

A neon sign reading 'Open' in English and the Chinese characters for open, on a brick wall at night.

It renders English and Chinese text.

Against its rivals

  • ChromaChroma is less precise on text but looser on style and content.
  • Stable Diffusion 3.5SD 3.5 is older and slower; Z-Image is newer.#37 of 37 on Artificial Analysis Text to Image Arena (840 Elo)
  • Hunyuan ImageHunyuan is much larger and follows long prompts better, at about nine times the price.#34 of 37 on Artificial Analysis Text to Image Arena (944 Elo)
Compare side by side ↗

Version history

Each version, dated, with what it changed over the one before.

  1. Z-Image-TurboLatest
    • Distilled 6B model with a single-stream DiT architecture and Decoupled-DMD distillation
    Source : huggingface.co

Benchmarks

How to read these scores: what a benchmark tells you.

Artificial Analysis Text to Image Arena: #35 of 37 models measured in the catalog. Thousands of people compare two images made from the same prompt without knowing the models, and pick the better one. An Elo rating, as in chess: higher is preferred more often. 100 points apart means the higher one wins about 64% of duels.

Artificial Analysis Text to Image ArenaElo rating · Artificial Analysis, October 4, 2026
  1. 1. GPT Image 2.5 SunburstGPT Image1,197
  2. 2. GPT Image 2.5 FlareGPT Image1,191
  3. 3. GPT Image 2GPT Image1,172
  4. 4. Nano Banana 2.1Nano Banana1,172
  5. 5. Grok Imagine 2.0Grok Imagine1,155
  6. 6. Nano Banana 2Nano Banana1,125
  7. 7. Muse ImageMuse Image1,115
  8. 8. GPT Image 1.5GPT Image1,107
  9. 9. Nano Banana ProNano Banana1,102
  10. 10. Nano Banana 2 LiteNano Banana1,096
  11. 35. Z-Image TurboZ-Image Turbo941
See the numbers
ModelElo
1. GPT Image 2.5 Sunburst1,197
2. GPT Image 2.5 Flare1,191
3. GPT Image 21,172
4. Nano Banana 2.11,172
5. Grok Imagine 2.01,155
6. Nano Banana 21,125
7. Muse Image1,115
8. GPT Image 1.51,107
9. Nano Banana Pro1,102
10. Nano Banana 2 Lite1,096
35. Z-Image Turbo941

The model in this family

ModelUsePrivacyPrice
Z-Image Turboz-image-turboImagePrivate$0.0045 per image

Specifications

SpecificationZ-Image-Turbo
ReleasedNovember 2025
Parameters6B
WeightsOpen, Apache 2.0
Acceptstext
Producesimages
LanguagesEnglish and Chinese text
Price$0.0045 per image
PrivacyPrivate

Prices, context and lengths from the live catalog on October 16, 2026; parameters and licenses from the maker.

Sources

Everything on this page comes from these pages, read on October 16, 2026 or as dated on each score.

Show the 3 sources
Learn more about Z-Image Turbo ↗