Wan 2.6 synchronous | Qwen Cloud

POST

/services/aigc/multimodal-generation/generation

curl --location 'https://dashscope-intl.aliyuncs.com/api/v1/services/aigc/multimodal-generation/generation' \
--header 'Content-Type: application/json' \
--header "Authorization: Bearer $DASHSCOPE_API_KEY" \
--data '{
  "model": "wan2.6-image",
  "input": {
    "messages": [
      {
        "role": "user",
        "content": [
          {
            "text": "Generate a tomato and egg stir-fry based on the style of image 1 and the background of image 2"
          },
          {
            "image": "https://cdn.wanx.aliyuncs.com/tmp/pressure/umbrella1.png"
          },
          {
            "image": "https://img.alicdn.com/imgextra/i3/O1CN01SfG4J41UYn9WNt4X1_!!6000000002530-49-tps-1696-960.webp"
          }
        ]
      }
    ]
  },
  "parameters": {
    "prompt_extend": true,
    "watermark": false,
    "n": 1,
    "enable_interleave": false,
    "size": "1K"
  }
}'

{
  "output": {
    "choices": [
      {
        "finish_reason": "stop",
        "message": {
          "content": [
            {
              "image": "https://dashscope-result-intl.oss-cn-singapore.aliyuncs.com/xxx.png?Expires=xxx",
              "type": "image"
            }
          ],
          "role": "assistant"
        }
      }
    ],
    "finished": true
  },
  "usage": {
    "image_count": 1,
    "input_tokens": 0,
    "output_tokens": 0,
    "size": "1376*768",
    "total_tokens": 0
  },
  "request_id": "a3f4befe-cacd-49c9-8298-xxxxxx"
}

Get an API key and set it as an environment variable. If using the SDK, install it.

This synchronous endpoint is for wan2.6-image only. Retrieve generated images directly in a single request without polling. For asynchronous processing, use the async submit endpoint.

Authorizations

string

header

required

DashScope API Key. Create one in the Qwen Cloud console.

Body

application/json

enum<string>

required

Model name. Set to wan2.6-image.

Available options:wan2.6-image

Example:wan2.6-image

object

required

Input data containing the messages array.

Show child attributes

object[]

required

Array of request content. Only single-turn conversations are supported. Provide one message with role: user.

Required range:items: 1–1

Show child attributes

enum<string>

required

Message role. Must be user.

Available options:user

object[]

required

Message content array. Must contain exactly one text object. Image objects depend on the mode:

Image editing (enable_interleave=false): 1 to 4 image objects required.
Interleaved text-image (enable_interleave=true): 0 to 1 image objects.

When using multiple images, include multiple image objects in the array. Image order is determined by array position.

Show child attributes

string

Positive prompt describing the desired image content, style, and composition. Supports Chinese and English. Maximum 2,000 characters (each Chinese character, letter, digit, or symbol counts as one character). Excess is auto-truncated. The content array must contain exactly one text object.

Example:Generate a tomato and egg stir-fry based on the style of image 1 and the background of image 2

Required range:length <= 2000

string

Input image as a public URL (HTTP/HTTPS) or Base64-encoded string (data:{mime_type};base64,{data}).

Image constraints:

Formats: JPEG, JPG, PNG (alpha channel not supported), BMP, WEBP.
Resolution: Width and height each between 240 and 8,000 pixels.
File size: Max 10 MB.

Image quantity limits:

When enable_interleave=false (image editing): must input 1 to 4 images.
When enable_interleave=true (interleaved text-image): can input 0 to 1 images.

Example:https://cdn.wanx.aliyuncs.com/tmp/pressure/umbrella1.png

object

Image processing parameters.

Show child attributes

string

Negative prompt describing content you do NOT want in the image. Supports Chinese and English. Maximum 500 characters. Excess is auto-truncated.

Example: Low resolution, low quality, deformed limbs, deformed fingers, oversaturated colors, wax-like appearance, no facial details, overly smooth skin, AI-looking artifacts, chaotic composition, blurry or distorted text.

Required range:length <= 500

string

Output image resolution. Supports two methods: referencing input image proportions or directly specifying dimensions.

Image editing mode (enable_interleave=false):

Method 1 (recommended): 1K (default) or 2K. Output total pixels close to 1280*1280 or 2048*2048, maintaining the aspect ratio of the last input image.
Method 2: Specify width*height in pixels. Total pixels must be between [768*768, 2048*2048], aspect ratio [1:4, 4:1]. Actual values are multiples of 16.

Interleaved text-image mode (enable_interleave=true):

Method 1 (default): References input image proportions. If total pixels <= 1280*1280, output matches input. If > 1280*1280, output scales to ~1280*1280.
Method 2: Specify width*height. Total pixels must be between [768*768, 1280*1280], aspect ratio [1:4, 4:1].

Recommended resolutions: 1280*1280 (1:1), 800*1200 (2:3), 1200*800 (3:2), 960*1280 (3:4), 1280*960 (4:3), 720*1280 (9:16), 1280*720 (16:9), 1344*576 (21:9).

Example:1K

boolean

defaultfalse

Controls image generation mode:

false (default): Image editing mode. Supports multi-image input (1-4 images) and subject consistency generation. Can generate 1 to 4 result images.
true: Interleaved text-image output mode. Supports 0-1 input images. Generates mixed content containing both text and images. Requires stream=true and X-DashScope-Sse: enable header.

integer

default4

Number of images to generate. Behavior depends on mode:

Image editing (enable_interleave=false): Range 1-4. Default: 4.
Interleaved text-image (enable_interleave=true): Must be 1. Use max_images to control image count instead.

Note: n directly affects cost. Cost = unit price x number of successfully generated images.

Required range:1 <= x <= 4

integer

default5

Only effective in interleaved text-image mode (enable_interleave=true). Specifies the maximum number of images the model can generate in a single response. Range: 1-5. Default: 5. The actual number of generated images is determined by model inference and may be less than this value.

Note: max_images affects cost. Cost = unit price x number of successfully generated images.

Required range:1 <= x <= 5

boolean

defaulttrue

Only effective in image editing mode (enable_interleave=false). Enables intelligent prompt rewriting that optimizes and refines the positive prompt. The negative prompt is not affected.

boolean

defaultfalse

Controls whether the response uses streaming output. In interleaved text-image mode (enable_interleave=true), you must set this to true.

boolean

defaultfalse

Adds a watermark label in the bottom-right corner of the image with fixed text "AI Generated".

integer

Random number seed. Range: [0, 2147483647]. Same seed produces more consistent (but not identical) results. If omitted, a random seed is used.

Required range:0 <= x <= 2147483647

Response

200-application/json

object

Show child attributes

object[]

List of generated results.

Show child attributes

string

stop means normal completion. null indicates streaming is still in progress.

Example:stop

object

Show child attributes

enum<string>

Always assistant.

Available options:assistant

object[]

Response content array containing generated image URLs and/or text (in interleaved mode).

Show child attributes

enum<string>

Content type: image or text.

Available options:image,text

string

URL of the generated image (PNG). Valid for 24 hours. Download promptly.

string

Text content (only present in interleaved text-image output mode).

boolean

Whether the generation has completed.

object

Usage statistics.

Show child attributes

integer

Number of images generated.

integer

Number of input tokens consumed.

integer

Number of output tokens consumed.

string

Resolution of the generated image(s).

integer

Total tokens consumed (input + output).

string

default"a3f4befe-cacd-49c9-8298-xxxxxx"

Unique request identifier for tracing and troubleshooting.

Example:a3f4befe-cacd-49c9-8298-xxxxxx