Baidu: ERNIE-Image-Turbo
ernie-image-turbo- Price
- From $0.0013
- Modalities
- Image
About ERNIE-Image-Turbo
ERNIE-Image-Turbo is Baidu's distilled, fast variant of ERNIE-Image, an open-weight text-to-image model. It produces an image in about 8 inference steps, optimized with DMD distillation and reinforcement learning, while keeping comparable quality to the full model in many cases. It suits rapid visual iteration and high-throughput ideation where speed and polished looks beat maximum instruction fidelity.
Where it works well
- Needs only about 8 inference steps against roughly 50 for the base model, so batches of drafts return much sooner.
- Keeps the family's strength with legible on-image text, useful for quick poster and title-card drafts.
- Distillation with reinforcement learning targets pleasing aesthetics, so first-pass images look finished.
- Takes plain text prompts, with the family's prompt enhancer expanding short ideas into fuller scene descriptions.
When to choose another model
- Baidu says it gives up some of the base model's stronger general capability and instruction fidelity in return for speed.
- Very dense or complicated layouts are safer on ERNIE-Image, which spends more steps on composition.
- Text-to-image only; it cannot edit an existing picture or follow a reference image.
Getting started
Create API key
Create a key in Console, then use it with every model on the platform.
Send your first request
Copy the example for your language and run it against the endpoint.
Text to imageTokenLab endpointPOST/v1/images/generationscurl -X POST "https://api.tokenlab.sh/v1/images/generations" \ -H "Authorization: Bearer sk-xxx" \ -H "Content-Type: application/json" \ -d '{ "operation": "text-to-image", "model": "ernie-image-turbo", "prompt": "A minimalist product photo of matte black headphones on a soft blue background." }'
Parameters
Text to image
| Parameter | Values and limits | Default | Notes |
|---|---|---|---|
height | ≤ 2048 | 1024 | Use an integer. Use multiples of 16. |
n | 1 | 1 | Use an integer. |
operation | Not provided | Not provided | Not provided |
output_format | jpeg, png, webp | Not provided | Default encoding is shared by these equivalent values: jpg, jpeg. |
prompt | Not provided | Not provided | Not provided |
seed | ≥ 0, ≤ 9007199254740991 | Not provided | Use an integer. |
steps | ≥ 2, ≤ 10 | 8 | Use an integer. |
width | ≤ 2048 | 1024 | Use an integer. Use multiples of 16. |
Fields outside this contract are rejected.
Pricing
The Verified price applies to Verified, which costs less on most models. The Official price is the model maker's published price and applies to the more reliable Official route. Auto bills the route that completes the request.
Text to image
per request- Official price
- $0.0013
- Verified price
- —
- Discount
- —
| Official priceper request | Verified priceper request | Discount | |
|---|---|---|---|
| Text to image | $0.0013 | — | — |
Usage & activity
Success rate is the share of requests that completed. Latency is how long a full response takes; P95 means 95% of requests finished within that time.
Usage & availability
Last 24 hours- Requests
- Success rate
- P95 latency
- Total tokens
Data is based on aggregate user requests, excluding status checks.
Open in Console
Open ERNIE-Image-Turbo in Console with a prompt ready to edit or send.
Help me create an image with ernie-image-turbo using text-to-image at /v1/images/generations. Show the result, cost, and any limits I should know.
Use cases
Best for- Text to Image
Rapid concept exploration
Generate dozens of visual directions for a campaign or product idea in one sitting, then promote the best prompt to the full model.
Social and blog graphics
Make title cards and illustrated headers quickly, with short text lines rendered inside the image.
Prompt testing
Try several prompt wordings quickly, then run the final version on a slower, higher-fidelity model.
Prompt examples
A flat-illustration title card reading 'WEEKLY RELEASE NOTES' on a teal background with small rocket icons.
A cozy bookstore window at dusk, warm lights, a handwritten sign that says 'Open Late'.
Six app icon concepts for a habit tracker in a soft gradient style, arranged in a grid.
FAQ
What is ERNIE-Image-Turbo and who makes it?
It is Baidu's distilled fast version of ERNIE-Image. It generates in about 8 steps rather than 50, using DMD distillation and reinforcement learning, and is meant for rapid iteration, ideation, and workflows that value speed and attractive output.
Is Turbo as good as the standard ERNIE-Image?
In many scenarios its quality is relatively comparable, according to Baidu. The base model keeps the edge on general-purpose capability and instruction fidelity, so use Turbo for drafts and exploration and the base model when a layout must match the prompt exactly.
Can ERNIE-Image-Turbo render text inside images?
Yes, text rendering is a trait of the ERNIE-Image family. For short titles and labels it works well for drafts. Proofread the spelling on every output, and switch to the base model when the text is long or tightly laid out.
What inputs does ERNIE-Image-Turbo accept?
A text prompt only. It does text-to-image generation, so there is no image upload, mask, or edit mode. Describe subject, style, and any text to appear in the image.
How much does ERNIE-Image-Turbo cost?
On TokenLab, ERNIE-Image-Turbo costs $0.0013 per request. The pricing table above shows the full breakdown. Rates depend on the billing unit, specification, and usage. Compare matching conditions in the model's detailed pricing; a single rate does not determine the total cost.
Which endpoint should ERNIE-Image-Turbo use?
Use https://api.tokenlab.sh/v1/images/generations for ERNIE-Image-Turbo. The request example below shows the matching code shape.
Which operations does ERNIE-Image-Turbo support?
ERNIE-Image-Turbo supports Text to image. Select an operation above to see its endpoint and request example.
Compare ERNIE-Image-Turbo
Sources
Reviewed Oct 2, 2026