> ## Documentation Index > Fetch the complete documentation index at: https://docs.powertokens.ai/llms.txt > Use this file to discover all available pages before exploring further. # Qwen text-to-image > Calls Ali `qwen-image-2.0-pro` model to perform text-to-image generation. The request body contains three main fields: `model`, `input`, and `parameters`. The prompt is placed in `input.messages[0].content[0].text`. The `parameters` field controls resolution `size`, count `n`, negative prompt `negative_prompt`, smart prompt rewriting `prompt_extend`, watermark `watermark`, and random seed `seed`. ## OpenAPI ````yaml api-reference/en/zmodelImage/ali/api/qwen-text-to-image.json POST /ali/api/v1/services/aigc/multimodal-generation/generation openapi: 3.0.1 info: title: Qwen text-to-image version: 1.0.0 description: >- Ali Qwen text-to-image capability documentation. The unified API endpoint is `/ali/api/v1/services/aigc/multimodal-generation/generation`. This document describes the public gateway contract for `qwen-image-2.0-pro`: `input.messages[].content[].text` carries the prompt, and `parameters` controls resolution, count, negative prompt, and other options. license: name: Project License url: https://github.com/QuantumNous/new-api/blob/main/LICENSE servers: - url: https://api.powertokens.ai description: Baze API server endpoint security: [] tags: - name: Qwen Image description: Qwen text-to-image capability paths: /ali/api/v1/services/aigc/multimodal-generation/generation: post: tags: - Qwen Image summary: Qwen text-to-image description: >- Calls Ali `qwen-image-2.0-pro` model to perform text-to-image generation. The request body contains three main fields: `model`, `input`, and `parameters`. The prompt is placed in `input.messages[0].content[0].text`. The `parameters` field controls resolution `size`, count `n`, negative prompt `negative_prompt`, smart prompt rewriting `prompt_extend`, watermark `watermark`, and random seed `seed`. operationId: aliQwenTextToImage requestBody: required: true content: application/json: schema: $ref: '#/components/schemas/TextToImageRequest' examples: qwen_image_2_pro_generation: summary: Qwen-Image-2.0 Pro text-to-image value: model: qwen-image-2.0-pro input: messages: - role: user content: - text: >- 冬日北京的都市街景，青灰瓦顶、朱红色外墙的两间相邻中式商铺比肩而立，檐下悬挂印有剪纸马的暖光灯笼，在阴天漫射光中投下柔和光晕，映照湿润鹅卵石路面泛起细腻反光。左侧为书法店：靛蓝色老旧的牌匾上以遒劲行书刻着“文字渲染”。店门口的玻璃上挂着一幅字，自上而下，用田英章硬笔写着“专业幻灯片 中英文海报 高级信息图”，落款印章为“1k token”朱砂印。店内的墙上，可以模糊的辨认有三幅竖排的书法作品，第一幅写着“阿里巴巴”，第二幅写着“通义千问”，第三幅写着“图像生成”。一位白发苍苍的老人背对着镜头观赏。右侧为花店，牌匾上以鲜花做成文字“真实质感”；店内多层花架陈列红玫瑰、粉洋牡丹和绿植，门上贴了一个圆形花边标识，标识上写着“2k resolution”，门口摆放了一个彩色霓虹灯，上面写着“细腻刻画 人物 自然 建筑”。两家店中间堆放了一个雪人，举了一老式小黑板，上面用粉笔字写着“Qwen-Image-2.0 正式发布”。街道左侧，年轻情侣依偎在一起，女孩是瘦脸，身穿米白色羊绒大衣，肉色光腿神器。女孩举着心形透明气球，气球印有白色的字：“生图编辑二合一”。里面有一个毛茸茸的卡皮巴拉玩偶。男孩身着剪裁合体的深灰色呢子外套，内搭浅色高领毛衣。街道右侧，一个后背上写着“更小模型，更快速度”的骑手疾驰而过。整条街光影交织、动静相宜。 parameters: negative_prompt: >- 低分辨率，低画质，肢体畸形，手指畸形，画面过饱和，蜡像感，人脸无细节，过度光滑，画面具有AI感。构图混乱。文字模糊，扭曲。 prompt_extend: true watermark: false size: 2048*2048 responses: '200': description: Success. Returns the image generation result. content: application/json: schema: $ref: '#/components/schemas/ImageResponse' examples: success: summary: Success response value: request_id: d0250a3d-b07f-49e1-bdc8-6793f4929xxx output: choices: - message: role: assistant content: - image: >- https://dashscope-result-sh.oss-cn-shanghai.aliyuncs.com/xxx.png?Expires=xxx finish_reason: stop usage: image_count: 1 height: 2048 width: 2048 '400': description: Invalid request parameters. content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' '401': description: Authentication failed, e.g. missing or invalid token. content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' '413': description: Request body too large. content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' '429': description: Rate limit triggered or insufficient account quota. content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' '500': description: Internal server error while processing the request. content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' '502': description: Upstream provider returned an abnormal response. content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' '503': description: >- Upstream provider is temporarily unavailable, or no available channel for the current model. content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' security: - BearerAuth: [] components: schemas: TextToImageRequest: type: object required: - model - input properties: model: type: string enum: - qwen-image-2.0-pro description: 'Model name. Example: `qwen-image-2.0-pro`.' input: $ref: '#/components/schemas/QwenTextToImageInput' parameters: $ref: '#/components/schemas/QwenTextToImageParameters' ImageResponse: type: object properties: output: $ref: '#/components/schemas/ImageOutput' description: Task output information. usage: $ref: '#/components/schemas/ImageUsage' request_id: type: string description: >- Unique request identifier. It can be used for request traceability and troubleshooting. code: type: string description: Error code. Only returned when the request fails. message: type: string description: Error message. Only returned when the request fails. ErrorResponse: type: object properties: code: type: string description: >- Error code returned when the request fails. This parameter is not returned on success. message: type: string description: >- Detailed error message returned when the request fails. This parameter is not returned on success. request_id: type: string description: >- Unique request identifier. It can be used for request traceability and troubleshooting. QwenTextToImageInput: type: object required: - messages properties: messages: type: array minItems: 1 maxItems: 1 description: >- Request content array. Currently only single-turn conversations are supported, and the array must contain exactly one element. items: $ref: '#/components/schemas/QwenTextToImageMessage' description: Basic input information. QwenTextToImageParameters: type: object properties: negative_prompt: type: string maxLength: 500 description: >- Negative prompt used to describe content that should not appear in the image and to constrain the image. Chinese and English are supported. Maximum length is 500 characters; extra content is automatically truncated. Example: low resolution, low quality, malformed limbs, malformed fingers, oversaturated image, wax figure appearance, no facial details, overly smooth, AI-like image, chaotic composition, blurry or distorted text. size: type: string description: >- Output image resolution in `width*height` format. For the `qwen-image-2.0` series, total output pixels must be between `512*512` and `2048*2048`, and the default resolution is `2048*2048`. Recommended resolutions: `2688*1536` (16:9), `1536*2688` (9:16), `2048*2048` (1:1, default), `2368*1728` (4:3), and `1728*2368` (3:4). 'n': type: integer minimum: 1 maximum: 6 default: 1 description: >- Number of output images. Default is 1. For the `qwen-image-2.0` series, you can generate 1-6 images. prompt_extend: type: boolean default: true description: >- Whether to enable smart Prompt rewriting. When enabled, the model optimizes and polishes the positive prompt. This feature does not modify the negative prompt. `true`: default, enabled. Recommended when you want more diverse image content and model-added details. `false`: disabled. Recommended when you want more controllable image details. Currently, only async APIs return the actual prompt. watermark: type: boolean default: false description: >- Whether to add a "Qwen-Image" watermark in the lower-right corner of the image. Default is `false`. seed: type: integer format: int64 minimum: 0 maximum: 2147483647 description: >- Random seed. Range [0, 2147483647]. Using the same `seed` value can make generated content relatively stable. If omitted, the algorithm automatically uses a random seed. Note: generation is probabilistic, so the same `seed` cannot guarantee identical results every time. description: Image processing parameters. ImageOutput: type: object properties: choices: type: array description: Generated output content. This array contains only one element. items: $ref: '#/components/schemas/ImageChoice' ImageUsage: type: object properties: image_count: type: integer description: Number of generated images. Currently fixed at 1. width: type: integer description: Generated image width in pixels. height: type: integer description: Generated image height in pixels. description: Output usage statistics. Only successful results are counted. QwenTextToImageMessage: type: object required: - role - content properties: role: type: string enum: - user description: Message role. This parameter must be set to `user`. content: type: array minItems: 1 maxItems: 1 description: >- Message content array. Only one `text` item is supported; omitting it or passing multiple text items causes an error. items: $ref: '#/components/schemas/QwenTextToImageContentItem' ImageChoice: type: object properties: finish_reason: type: string description: Reason the task stopped. `stop` indicates a natural stop. message: $ref: '#/components/schemas/ImageChoiceMessage' QwenTextToImageContentItem: type: object required: - text properties: text: type: string description: >- Positive prompt used to describe the image content, style, and composition you want to generate. Chinese and English are supported. The `qwen-image-2.0` series supports up to 1300 tokens; other models support up to 800 tokens. Extra content is automatically truncated. ImageChoiceMessage: type: object properties: role: type: string description: Message role. Fixed to `assistant`. content: type: array description: Message content array returned by the model. items: $ref: '#/components/schemas/ImageChoiceContentItem' description: Message returned by the model. ImageChoiceContentItem: type: object properties: image: type: string description: >- Generated image URL. The image format is PNG. The link is valid for 24 hours; download and save the image promptly. securitySchemes: BearerAuth: type: http scheme: bearer bearerFormat: JWT description: 'Pass `Authorization: Bearer ` in the request header.' ````