Skip to content
Kenerate IconKenerate Icon
Kenerate AI
Loading your account

Alibaba's Tongyi Lab (Tongyi-MAI)

Z-Image — Alibaba Tongyi's fast 6B image model

Z-Image is a 6-billion-parameter open image model from Alibaba's Tongyi Lab. On Kenerate, Z Image Base remixes an uploaded image with a strength slider.

Architecture
6B-parameter single-stream diffusion transformer
Released
Z-Image-Turbo November 2025; Z-Image January 2026
Weights
Open (Apache 2.0)
Maker's highlights
Photorealism, English and Chinese text rendering, instruction following, fits 16 GB consumer GPUs
On Kenerate
Z Image Base — up to 2048 px, one reference with a strength slider
Try
  • Z-Image
  • Kitchen portrait
  • Neon sign with text
  • Night market
  • Sleeping cat
  • Real outputs
  • Official sources
  • Free to start
  • No watermark

Updated by the

Made with Kenerate

What is Z-Image?

Updated

What is Z-Image?

Z-Image is an open image model family from Alibaba's Tongyi Lab, built on a single-stream diffusion transformer for photoreal images and released from November 2025. On Kenerate AI, Z Image Base makes cheap photoreal drafts and remixes one uploaded image with a strength slider — higher strength changes more.

Key facts

Maker
Alibaba Tongyi Lab
Size
6B parameters
Weights
Open · Apache 2.0
Z-Image-Turbo
Nov 2025
Z-Image
Jan 2026
On Kenerate
Z Image Base · 2048 px
Image to image
1 reference · strength slider
Runs on
16 GB consumer GPUs

Gallery

Z-Image examples

Real outputs made on Kenerate. Open any image to copy its prompt and run it yourself.

Same prompt

Z Image Base strength test: 35 %, 60 %, 85 %

The same pencil sketch and prompt at three Reference strength settings. Higher means more change.

PromptFinished fantasy oil painting of a warrior woman holding a spear with a flowing crimson cape, dramatic sunset sky, rich colors

  • Your image: Finished fantasy oil painting of a warrior woman holding a spear with a flowing crimson cape, dramatic sunset sky, rich colorsYour imagePencil sketch
  • 35 % strength: Finished fantasy oil painting of a warrior woman holding a spear with a flowing crimson cape, dramatic sunset sky, rich colors35 % strengthCleaner sketch
  • 60 % strength: Finished fantasy oil painting of a warrior woman holding a spear with a flowing crimson cape, dramatic sunset sky, rich colors60 % strengthInked drawing
  • 85 % strength: Finished fantasy oil painting of a warrior woman holding a spear with a flowing crimson cape, dramatic sunset sky, rich colors85 % strengthFull-colour painting

Models

Z-Image on Kenerate

Every version in the generator, and what each is good at.

  • Z Image Base title card — a portrait of a red fox that transitions smoothly from photograph on the left to oil painting on the right

    Alibaba Tongyi Lab

    Z Image Base

    Reference images
    1 (with strength)
    Output
    Up to 2048 px

    Best for

    Cheap photoreal drafts; image remix

    Open Z Image Base

From our tests

Strengths and limits

Good at

  • Photoreal faces and textures at a low price — the kitchen portrait and sleeping cat are typical.
  • Rendered the neon 'OPEN LATE' sign correctly on the first try.
  • A strength slider that really works for image to image: 35 % keeps your image, 85 % repaints it.
  • Up to 2048 px.

Limits

  • Leans to an anime look on some fantasy prompts (see the knight) — add “photorealistic” to steer it.
  • One reference image only; for combining images use Seedream 5.0 Pro, Nano Banana 2 or GPT Image 2.
  • Not on the free plan — any credit pack unlocks it.

How to

How to use Z-Image on Kenerate

  1. Step 1

    Open Z Image Base

    Pick it in /app/image. Any credit pack unlocks it.

  2. Step 2

    Write the prompt

    Plain sentences work; say “photorealistic” for photos.

  3. Step 3

    Remix an image (optional)

    Upload one image and set Reference strength in Advanced settings — lower keeps more.

  4. Step 4

    Download

    Up to 2048 px, no watermark.

1Describe your image2Pick an image model3Set the aspect ratio and size4Click Create5Open and download your image
Kenerate AI Image Studio with a text prompt typed into the prompt box

Open Image Studio and type what you want to see — subject, style, lighting and mood.

1 / 5 · Describe your image

Watch the tour · 1:23

Prompts, styles, references, edit, Arena & chat

Chapters

Example uses

How creators use Z-Image

Illustrative examples of typical workflows, not customer reviews.

  • Fantasy painting of a misty forest with sunbeams and a small deer

    Concept artist

    A concept artist could upload a rough pencil sketch and raise the strength slider until it becomes a painted scene, using lower settings when only cleaner line work is needed.

  • Ginger cat asleep on a pile of books by a sunny window

    Pet blogger

    A pet blogger could make cheap photoreal drafts of cosy animal scenes for posts before choosing a final image.

  • Lantern-lit night market street with food stalls and steam

    Travel blogger

    A travel blogger could draft atmospheric street scenes at up to 2048 px, adding 'photorealistic' to keep the look true to life.

FAQ

Z-Image questions

What is Z-Image?

Z-Image is an open 6-billion-parameter image model family from Alibaba's Tongyi Lab, built on a single-stream diffusion transformer. Z-Image-Turbo was released in November 2025 and Z-Image in January 2026. Its maker highlights photorealism, English and Chinese text rendering and instruction following, and says it fits 16 GB consumer GPUs. On Kenerate it runs as Z Image Base, up to 2048 px.

Link to this answer

How much does Z-Image cost on Kenerate?

It is credit-based, and the cost of each image is shown before you generate. Credits come in one-time packs that never expire, with no subscription required, and current packs are on /pricing. Z-Image is not on the free plan, but any credit pack unlocks it, and it is one of the cheaper ways to make photoreal drafts on Kenerate.

Link to this answer

Is Z-Image open source?

Yes. Tongyi Lab released the Z-Image weights under the Apache 2.0 licence, so you can download and run the model yourself. Its single-stream design at 6 billion parameters keeps it small enough for 16 GB consumer GPUs. If you would rather not set it up locally, Z Image Base runs in the browser on Kenerate, up to 2048 px with no watermark.

Link to this answer

How does the strength slider work on Z Image Base?

It appears under Advanced settings when you upload an image, from 10 to 100 %, with 75 % as the default. Lower values keep more of your image; higher values change more. In our test with a pencil sketch, 35 % gave a cleaner sketch, 60 % inked it, and 85 % turned it into a full-colour painting. Z Image Base takes one reference image.

Link to this answer

Can Z-Image write text?

Short text, yes. Our neon 'OPEN LATE' sign came out correctly on the first try, and its maker highlights English and Chinese text rendering. Keep each text block short and put the exact words in quotes. For long text, posters or text-heavy designs, GPT Image 2 is stronger, so switch models for those jobs.

Link to this answer

In depth

Z-Image, in depth

Background, how to prompt it, and when to pick something else.

What makes Z-Image different

Most large image models use two streams for text and image. Z-Image uses a single-stream diffusion transformer at 6 billion parameters, which keeps it small enough for 16 GB consumer GPUs while producing photoreal results [1].

Keep going

More models

By Last updated and reviewed

We run these models in the Kenerate generator. Model facts come from the makers' own pages; specs and limits were checked against our code; the examples are our own outputs.

Sources: Tongyi-MAI — Z-Image (GitHub)

Try Z-Image

Free to start. No watermark.

Try Z-Image

Tried it? Tell us how it went