IllustrationStudio product shot
Text to image · 1:1
Prompt: A shiny red apple with a green leaf on a wooden table, plain light background, soft studio light, product photo
Z-Image AI Image Generator
Alibaba’s fast, open-weight Z-Image Turbo, ready to use online. At 3 credits an image, make four at a time and keep the best.
Z-Image prompts to start from
Short, plain prompts work well on Z-Image. Click one to load it into the generator, then run four and pick your favorite.
Illustration
IllustrationCharacter portrait
Text to image · 1:1
Prompt: A portrait of a smiling elderly fisherman with a gray beard and a black knit cap, yellow rain jacket, harbor and boats behind him at sunrise, photo
IllustrationLandscape at sunrise
Text to image · 1:1
Prompt: A mountain lake at sunrise with mist on the water, snowy peaks and pine trees reflected in the lake, rocks in the foreground, landscape photo
IllustrationCute 3D illustration
Text to image · 1:1
Prompt: A cute white robot watering pink flowers on a sunny balcony garden, city skyline in the background, soft 3D illustration
What is Z-Image?
Z-Image is an open-weight image model from Tongyi-MAI, an Alibaba lab. Turbo is its fast version, built to make an image in a few sampling steps.
On Imgveo, Z-Image Turbo costs 3 credits an image, our lowest price for a paid image model. It makes 1 to 4 images per run from a text prompt, in 5 aspect ratios.
Tongyi-MAI describes a 6-billion-parameter model that renders English and Chinese text, released under the Apache 2.0 license. Turbo came out in November 2025.
Z-Image on Imgveo today
What you can make right now. Every number comes from the generator’s own settings.
- Version
- Z-Image Turbo
- Starts from
- A text promptNo reference images or edits
- Text in images
- English and Chinese
- Aspect ratios
- 1:1 · 4:3 · 3:4 · 16:9 · 9:16
- Images per run
- 1 to 4
- Price
- 3 credits per imageFour images: 12 credits
Turbo, Base, Edit: the Z-Image family
Tongyi-MAI lists four versions. Only one of them is what this page runs.
| Version | Sampling steps | What it does | On Imgveo |
|---|---|---|---|
| Z-Image Turbo | 8 | Fast text-to-image | Yes, on this page |
| Z-Image | 50 | Text-to-image, slower and tunable | No |
| Z-Image Edit | 50 | Edits an existing image | Not released yet |
| Z-Image Omni-Base | 50 | Generation and editing in one model | Not released yet |
Release status as listed by Tongyi-MAI when we checked in November 2026. Turbo’s speed comes from distillation: fewer steps, so each image comes back fast.
Run it here or on your own GPU?
Z-Image’s weights are open, so you can run it yourself. Here is what each route takes.
Here on Imgveo
- Nothing to install, works on any device
- 3 credits an image, four per run
- Results saved to your library
- Failed images are refunded
On your own machine
- A GPU with about 16 GB of memory, per Tongyi-MAI
- ComfyUI or a Python setup, and the model download
- No per-image cost once it runs
- Best if you want to fine-tune or automate
How to use Z-Image in 3 steps
Quick enough to run again and again until it’s right.
01
Describe the picture
Subject, setting, style. Plain words work; you don’t need long lists of quality tags.
02
Pick a shape and 1 or 4 images
5 aspect ratios. Four images cost 12 credits and give you choices.
03
Keep the best one
Download the one you like, tweak the prompt and run again. Failed images are refunded.
English and Chinese text
Tongyi-MAI built Z-Image to render text in both languages. Short words on signs and labels are its sweet spot.
IllustrationEnglish neon sign
Prompt
A neon sign that reads "OPEN 24H" on a dark brick wall at night, pink letters in a blue neon frame, ivy at the top
IllustrationChinese shop banner
Prompt
A bowl of beef noodle soup on a wooden table, chopsticks lifting noodles, steam rising, a red shop banner that reads "面馆" in the background, red lanterns
01
Quote the exact words
Write “OPEN 24H” or “面馆” in quotes, exactly as it should appear.
02
Keep it short
A word or a short line on a sign, cup or banner. Long paragraphs belong on a text-first model like Qwen Image 3.0.
03
Write Chinese in characters
Use the characters themselves, not pinyin or a translation.
04
Check at full size
Zoom in before you use the image. At this price, a re-run is cheaper than a fix.
Cheap enough to make four and pick one
Product shots on plain backgrounds, flat lays and social images are where a low-cost model pays off. Each example pairs an illustration of the idea with a prompt written for Z-Image.
IllustrationProduct on white
Prompt
A white and gray running sneaker with blue suede details on a plain white background, side view, product photo
IllustrationDesk flat lay
Prompt
A flat lay of a desk with a laptop, a notebook, a coffee cup, headphones and a plant, plain light background, top view
The Z-Image prompt formula
Four short parts. Z-Image does well with clear, simple sentences.
- 01
Subject
What the image is of: a red apple, an old fisherman, a robot.
- 02
Setting
Where it is: a wooden table, a harbor at sunrise, a sunny balcony.
- 03
Style
Photo, 3D illustration, watercolor, flat vector.
- 04
Details
Light, angle and background: soft window light, top view, plain white background.
Copy-ready templates
Fill in the brackets and run four.
Product shot
Text to image · 1:1
A [product] on [surface or plain background], [lighting], [angle], product photo
Portrait
Text to image · 3:4
A portrait of [person: age, look, clothing], [expression], [setting] behind them, [light], photo
Landscape
Text to image · 16:9
[Place] at [time of day], [weather or mist], [foreground detail], [color mood], landscape photo
Illustration
Text to image · 1:1
A [cute / flat / watercolor] illustration of [subject] [doing what] in [setting], [palette]
Three common mistakes
01
Expecting edits
Z-Image here takes no reference image. To change a photo, use Seedream 5.5 or Nano Banana 2 Lite.
02
Expecting 4K
There is no resolution setting. For large prints, pick a model with 2K or 4K output.
03
Long text in the picture
Keep words to a short line. Menus and infographics need a text-first model.
Z-Image pricing
Every number below comes from the same code that charges you, so it changes when our prices do.
How the price works
3 credits per image, whatever the shape. One run makes up to four: 12 credits.
What a plan gets you
Starter’s 1,500 monthly credits make 500 Z-Image images.
Failed images are refunded
If an image fails, its credits go back to your balance automatically.
When to step up from Z-Image
Low-cost image models on Imgveo, priced by our generator. Move up a rung when you need what the next one adds.
| Model | Credits per image | Reference images | Resolution options |
|---|---|---|---|
| Z-Image Turbo | 3 | No | Fixed |
| Seedream 5 Flash | 5 | Yes | 1K / 2K |
| Grok Imagine Image 2.0 | 6 | Yes | Fixed |
| Nano Banana 2 Lite | 6 | Yes | Fixed |
| Qwen Image 3.0 | 8 | No | 1K / 2K |
| GPT Image 2.5 | 10 | Yes | 1K / 2K / 4K |
On the free plan? Flux 2 Klein is included and runs on your free monthly credits.
What people make with Z-Image
Everyday images where speed and price matter more than 4K.
Concept drafts
Try ten directions for a campaign before you spend on a premium model.
Social posts
Daily images for feeds and stories, four options per run.
Product backgrounds
Clean product-style shots on plain or simple backgrounds.
Blog and newsletter images
Header images that fit the topic, without stock-photo fees.
Mood boards
Fill a board with styles, colors and scenes in minutes.
Short bilingual text
Signs, cups and banners with a few words in English or Chinese.
Z-Image FAQ
Want every AI model in one place?
Open Imgveo Studio to run this and every other AI video and image model from a single workspace, with one credit balance.
Related models
Other image models on Imgveo to try next.
Make four images in one quick run
Describe it, run four, keep the best. Failed images are refunded.
Z-Image is developed by Tongyi-MAI at Alibaba. Imgveo is an independent platform and is not affiliated with Alibaba.