Skip to main content
QwenCloud home page
Docs
Qwen

Qwen — OpenAI-compatible image generation

Call the qwen-image-3.0 series for text-to-image and image-to-image/editing through the OpenAI Images-compatible protocol

POST
/compatible-mode/v1/images/generations
curl --location 'https://maas.qwencloudapi.com/compatible-mode/v1/images/generations' \
--header 'Content-Type: application/json' \
--header "Authorization: Bearer $DASHSCOPE_API_KEY" \
--data '{
  "model": "qwen-image-3.0-pro",
  "prompt": "A vertical outdoor portrait photograph with a warm afternoon street atmosphere. Deep green vines and small orange flowers cascade from building eaves across the upper area. A dark blue sign reads '\''Il Messaggero'\'' in white Gothic lettering, partially obscured by foliage. Below, a newsstand displays newspapers behind black metal-framed glass, blurred by shallow depth of field. Strong backlight streams from the street'\''s end. Center-right, a young woman in a black spaghetti-strap backless dress looks back at the camera with a warm smile. Her long, thick wavy black hair is outlined by golden rim light. She has fair skin, bright eyes, soft coral-red lips, and holds a large bouquet of orange, apricot, pink and peach roses contrasting with her black dress. The sunlit city street stretches into the blurred background. Warm film-like tones with fine grain, soft contrast and pronounced backlit edge glow create a romantic, bright, urban strolling atmosphere.",
  "size": "1024x1024",
  "n": 1,
  "prompt_extend": true
}'
{
  "created": 1788339600,
  "data": [
    {
      "url": "https://dashscope-result-sg.oss-ap-southeast-1.aliyuncs.com/xxx.png?Expires=xxx"
    }
  ],
  "usage": {
    "output_height": 1024,
    "output_width": 1024,
    "input_image_count": 0,
    "input_image_type": "qima_input_1k",
    "output_image_count": 1,
    "output_image_type": "qima_output_1k"
  }
}
Get an API key and set it as an environment variable. If using the SDK, install it.
Image generation can take a while. Set a generous client-side timeout. For multi-image output (larger n) or concurrent calls, start at 600 seconds to avoid the client disconnecting early.

Authorizations

​
string
header
required

DashScope API key. Get your API key from QwenCloud console.

Body

application/json
​
enum<string>
required

Model name.

Available options:qwen-image-3.0-pro,qwen-image-3.0
Example:qwen-image-3.0-pro
​
string
required

Positive prompt describing the desired image content, style, and composition. Chinese and English supported, up to ~4500 tokens. Cannot be empty.

Example:A vertical outdoor portrait photograph with a warm afternoon street atmosphere. Deep green vines and small orange flowers cascade from building eaves across the upper area. A dark blue sign reads 'Il Messaggero' in white Gothic lettering, partially obscured by foliage. Below, a newsstand displays newspapers behind black metal-framed glass, blurred by shallow depth of field. Strong backlight streams from the street's end. Center-right, a young woman in a black spaghetti-strap backless dress looks back at the camera with a warm smile. Her long, thick wavy black hair is outlined by golden rim light. She has fair skin, bright eyes, soft coral-red lips, and holds a large bouquet of orange, apricot, pink and peach roses contrasting with her black dress. The sunlit city street stretches into the blurred background. Warm film-like tones with fine grain, soft contrast and pronounced backlit edge glow create a romantic, bright, urban strolling atmosphere.
​
string

URL or Base64 encoded data of the input image. Omit for T2I; provide for I2I (1-3 images). Pass a string for one image, or an array of strings for multiple (array order defines image order). Do not pass null or an empty array, or a 400 is returned. Formats: JPG/JPEG/PNG/BMP/TIFF/WEBP/GIF; width and height 384-2048 px recommended; size up to 10MB.

Supported input formats:

  1. Public URL: HTTP and HTTPS protocols are supported.
  2. Base64 encoding: Format is data:{MIME_type};base64,{base64_data}.
​
integer
default1

Number of output images. A string (e.g. "1") returns 400; must be an integer.

Required range:1 <= x <= 6
​
string
default"auto"

Output resolution as widthxheight (letter x separator). auto lets the model recommend. Pixel area 512x512 to 2048x2048, aspect ratio 1:8 to 8:1. Note this differs from the DashScope * separator; adjust when migrating.

Example:1024x1024
​
string

Negative prompt describing content you do not want in the image.

​
integer

Random seed. Random when omitted; a fixed seed makes results relatively stable.

Required range:0 <= x <= 2147483647
​
boolean
defaulttrue

Whether to enable intelligent prompt rewriting (recommended). Rewrites the positive prompt per prompt_extend_mode.

​
enum<string>
default"direct"

Prompt rewriting mode. direct: direct prompt enhancement (DPE), supports both T2I and I2I; agent: agentic prompt enhancement (APE), T2I only; passing it for I2I returns 400.

Available options:direct,agent
​
boolean
defaulttrue

Whether to enable thinking mode. Improves image quality with stronger reasoning but increases latency. Effective only when prompt_extend=true; I2I Agent is not yet supported.

​
boolean
defaultfalse

Whether to add a watermark.

Response

200-application/json
​
integer

Unix timestamp (seconds) when the response was created.

​
object[]

List of generated results. Contains multiple elements when n > 1.

​
object

Resource usage for this call, returned only on success. This is image input/output metering, not token usage.

​
object

Error info, returned only on failure. OpenAI-compatible mode does not return request_id in the body; the request id is returned via the x-request-id HTTP header. With the OpenAI Python SDK, read response._request_id on success or APIStatusError.request_id on failure.