Qwen Image 2.0

Alibaba Cloud · Image Generation

POST /v1/images/generations

Unified image generation and editing model with class-leading complex Chinese/English text rendering, realistic textures, and multi-image fusion.

At a glance

Field	Value
Model id	`qwen-image-2-0`
Input modalities	Text, Image
Output modalities	Image
Context window	-
Weight precision	-
Region	Singapore
Features	text_rendering, image_editing, multi_image
Native inference	No
New	No
Supported endpoints	`POST /v1/images/generations`

Pricing

Charge	Spec	Rate
Standard	per image	$0.0322 (was $0.035)
Pro	per image	$0.069 (was $0.075)

Example request

$ curl https://api.empiriolabs.ai/v1/images/generations \
>   -H 'Authorization: Bearer $EMPIRIOLABS_API_KEY' \
>   -H 'Content-Type: application/json' \
>   -d '{"model": "qwen-image-2-0", "prompt": "a misty forest at dawn", "size": "1024x1024"}'

Parameters

Parameter	Type	Required	Default	Description
`prompt`	string	yes	-	Text prompt. With input images attached, becomes an editing instruction.
`model_variant`	enum	no	`"qwen-image-2.0-pro-2026-04-22"`	Pro variants offer higher quality. Dated snapshots pin a specific weights revision for reproducibility; the undated “qwen-image-2.0-pro” tracks the latest pro release. · Allowed: `qwen-image-2.0-pro-2026-04-22`, `qwen-image-2.0-pro-2026-03-03`, `qwen-image-2.0`
`aspect_ratio`	enum	no	-	Optional. If omitted, text-to-image defaults to 16:9; image-edit matches the last input image’s native ratio. · Allowed: `16:9`, `9:16`, `1:1`, `3:2`, `2:3`, `4:3`, `3:4`, `5:4`, `4:5`
`use_custom_size`	boolean	no	false	Override aspect_ratio with explicit pixel dimensions.
`custom_width`	number	no	-	Pixels (512–2048). Total pixel area must be in [512×512, 2048×2048]. Only when use_custom_size=true. · Range: 512 – 2048
`custom_height`	number	no	-	Pixels (512–2048). Total pixel area must be in [512×512, 2048×2048]. Only when use_custom_size=true. · Range: 512 – 2048
`image`	array	no	-	Up to 3 input image URLs. Triggers image-edit mode.
`num_images`	number	no	`1`	Number of images to generate. Each is billed separately. · Range: 1 – 6
`negative_prompt`	string	no	`""`	What to avoid in the output.
`prompt_extend`	boolean	no	true	Let DashScope rewrite the prompt for better results.
`watermark`	boolean	no	false	Forwarded to DashScope as parameters.watermark. When true, the upstream Qwen-Image service stamps a small Qwen provenance watermark on each generated image. EmpirioLabs does not add any watermark of its own.
`seed`	number	no	-	Reproducibility seed. Leave empty for random. · Range: 0 – 2147483647
`response_format`	enum	no	`"url"`	url (default) returns a signed URL; b64_json returns inline base64. · Allowed: `url`, `b64_json`

Notes

Modes

Text-to-image: no images attached
Image-editing: upload 1-3 images

Constraints

Positive prompt: max 800 characters
Negative prompt: max 500 characters
Custom width/height: 512-2048 px (total area between 512x512 and 2048x2048)
Per image is billed

Performance

Responses take 10-30 seconds depending on resolution, image count, and prompt complexity

Machine-readable schema: GET https://api.empiriolabs.ai/v1/models/qwen-image-2-0.