Image generation

Bailian Qwen-Image series

Generate images with Tongyi Qwen Bailian Qwen-Image models. Supports prompt refinement and strong Chinese-style semantics.

Basics

Endpoint: POST /v1/images/generations
Auth: Bearer Token
Content-Type: application/json

Model overview

Model	Summary	Output
wan2.6-t2i `recommended`	Wanxiang 2.6—free size within total pixel and aspect constraints (same idea as 2.5)	Resolution: total pixels in [1280×1280, 1440×1440]; aspect [1:4, 4:1]; PNG
wan2.5-t2i-preview `recommended`	Wanxiang 2.5 preview—flexible sizing within pixel/aspect bounds (e.g. 768×2700; 2.2 and below cap single side at 1400)
wan2.2-t2i-flash	Wanxiang 2.2 flash—~50% faster than 2.1	Resolution: both sides in [512, 1440]; PNG
wan2.2-t2i-plus	Wanxiang 2.2 plus—better stability vs 2.1
wanx2.1-t2i-turbo	Wanxiang 2.1 flash
wanx2.1-t2i-plus	Wanxiang 2.1 plus
wanx2.0-t2i-turbo	Wanxiang 2.0 flash

Request body

Field	Type	Required	Description
model	`string`	Yes	e.g. `qwen-image-plus` or `qwen-image-max`.
content	`string`	Yes	Image description text.
parameters	`object`	No	Generation controls.
└─ negative_prompt	`string`	No	Things to avoid.
└─ prompt_extend	`boolean`	No	AI prompt expansion (default true).
└─ watermark	`boolean`	No	Add watermark or not.
└─ size	`string`	No	e.g. `13281328`, `10241024`; width and height each 512–1440.
└─ n	`integer`	No	Number of images, 1–4, default `4`. Billed per image—use `1` for tests.
└─ seed	`integer`	No	Seed in `[0, 2147483647]`. Same seed tends to stabilize output but is not a hard guarantee.

Response

200 OK

Field	Type	Description
created	`integer`	Unix timestamp.
data	`array`	Results.
└─ url	`string`	Download URL.
└─ b64_json	`string`	Base64 if requested.
└─ revised_prompt	`string`	Prompt after AI refinement.

Request example

{
    "model": "wan2.2-t2i-flash",
    "prompt": "A kitten running on grass, cinematic look, natural light and shadow",
    "parameters":{
        "size": "960*1390"
    }
}

Response example

{
  "created": 1713833628,
  "data": [
    {
      "url": "https://example.com/qwen_output_image.png",
     	"b64_json": "",
      "revised_prompt": ""
    }
  ]
}

Tips

Chinese semantics: Qwen is strong on idioms, couplets, and Chinese motifs (e.g. Yueyang Tower, blue-and-white porcelain).
Prompt extend: Keeping prompt_extend: true usually yields richer detail.

Basics

Relay overview

Images

Video

Chat

Privacy and terms

Bailian Qwen-Image series

Basics

Model overview

Request body

Response

Request example

Response example

Tips

​Bailian Qwen-Image series

​Basics

​Model overview

​Request body

​Response

​Request example

​Response example

​Tips

Bailian Qwen-Image series

Basics

Model overview

Request body

Response

Request example

Response example

Tips