还没有图片。提交表单以生成图片。
探索不同用例和参数配置
透明定价,无隐藏费用。按需支付。
| 规则与模式 | 渠道 | 积分 | 价格(美元) | 官方/参考价格 | 每日节省 |
|---|---|---|---|---|---|
test imageQwen | Regular | 999per run | $4.541 | - | - |
qwen-image-2-1 使用完整指南
在 ApiPass 上使用 Qwen-Image-2.1 构建应用,通过一个对开发者友好的 API 实现文本生成图像、图像到图像编辑、透明素材、多参考图工作流和局部编辑。Qwen Image 2.1 是一款紧凑高效的图像模型,专为高质量创意与生产工作流而设计。

Qwen Image 2.1 是 Alibaba 推出的紧凑型 AI 图像模型,用于统一的图像生成与编辑。它支持文本生成图像、基于参考图的图像编辑、原生透明背景输出、最多 10 张图像的多参考合成、基于 mask 的局部编辑,并提升了人物、产品和文字的保真度。通过 ApiPass,开发者只需使用一个模型名称 qwen/qwen-image-2-1;当未提供参考图像时,适配器会自动选择文本生成图像路由,而当 image_urls 包含图像时,则自动选择图像到图像路由。
根据自然语言提示词生成精美视觉内容。Qwen Image 2.1 可创建人像、产品视觉图、插画、营销图形、社交媒体内容、海报及其他创意素材,具备出色的细节表现与审美效果。
使用文本指令改造现有图像。可更改物体、服装、表情、背景、风格和场景细节,并在可能的情况下保留源图像的视觉结构。
使用最多 10 张参考图进行多素材合成、虚拟试穿、产品摆放或场景搭建。对于定向修改,可提供 mask_url 并搭配且仅搭配一张参考图,只重新生成白色 mask 区域。
以 1K 或原生 2K 分辨率创建细节丰富的视觉内容,并支持多种宽高比。2K 生成非常适合产品视觉图、营销素材、海报、社交内容以及可用于生产的创意工作流。
无需依赖单独的背景移除步骤,即可直接生成透明视觉素材。透明背景输出使用带 alpha 通道的 PNG 或 WebP,且不能与 JPEG/JPG 输出同时使用。
从多个视觉输入创建一个连贯的结果。可将人物合成为合照,匹配服装和配饰,摆放产品,或由独立素材构建室内场景。
使用 mask_url 控制编辑发生的位置。白色 mask 区域会被重新生成,黑色区域保持不变。mask 编辑必须且只能与一张参考图一起使用,并且不能与透明背景同时使用。
Qwen Image 2.1 提升了人像的身份一致性,能保留产品的关键细节,并能更有效地处理海报、广告、信息图和社交图形等文字密集型设计。
构建文本生成图像产品,让用户通过自然语言提示词创建人像、插画、概念艺术、营销视觉图和社交内容。
创建图像编辑器,让用户通过简单的文本指令更改物体、背景、服装、表情、风格或选定区域。
生成营销活动视觉图,更新产品背景,创建目录内容,并在编辑过程中保留形状、纹理和文字等重要产品细节。
使用多参考图输入,将人物与服装、鞋子、包袋、帽子或配饰组合成统一的时尚视觉图,用于购物和造型体验。
直接从提示词或源图像创建用于产品、贴纸、设计元素和可复用创意内容的透明 PNG 或 WebP 素材。
ApiPass 提供了一种对开发者友好的方式来访问 Qwen Image 2.1,无需管理模型托管、GPU 基础设施或单独的上游路由。
使用 qwen/qwen-image-2-1 同时处理文本生成图像和图像到图像工作流。ApiPass 会根据 image_urls 是否包含参考图像自动路由请求。
使用 /api/v1/jobs/createTask 创建任务,然后使用返回的 taskId 通过 /api/v1/jobs/recordInfo 查询进度和结果。如果提供 callBackUrl,ApiPass 会向你的 endpoint 发送完成通知。
使用可选的 channel 参数选择 auto、starter、regular 或 official 路由。auto 会让 ApiPass 根据实时价格和稳定性,在可用服务商之间分配任务。
Qwen Image 2.1 支持 1K 和 2K 输出分辨率档位。2K 可以生成更精细的图像,但生成时间可能比 1K 更长。
ApiPass 旨在为图像生成 API 提供灵活且具成本意识的访问方式,让团队无需管理昂贵基础设施,也能测试想法并扩展生产工作负载。
新用户可使用免费额度,在投入生产前测试提示词、图像输入、编辑工作流、透明背景和结果质量。
使用 ApiPass Playground 测试提示词、上传受支持的参考图像、比较生成和编辑工作流,并直接在浏览器中查看生成的图像。
创建 ApiPass API key 用于身份验证。调用 Qwen Image 2.1 endpoints 时,请在 Authorization: Bearer YOUR_API_KEY header 中使用它。
向 /api/v1/jobs/createTask 发送 POST 请求,使用模型 qwen/qwen-image-2-1,并提供包含 prompt、可选 image_urls、mask_url、aspect_ratio、resolution、background、output_format、enhance_prompt 和 seed 的 input 对象。
使用返回的 taskId 查询 /api/v1/jobs/recordInfo。对于异步集成,请在创建任务时提供 callBackUrl,任务达到终止状态时 ApiPass 会通知你的 endpoint。
所有 API 都需要通过 Bearer Token 进行身份验证。
Authorization: Bearer
Submit a new Qwen Image 2.1 text-to-image or image-to-image generation task
The API accepts a JSON payload with the following structure:
1{
2 "model": "qwen/qwen-image-2-1",
3 "callBackUrl": "string (optional)",
4 "channel": "auto",
5 "input": {
6 "prompt": "string",
7 "image_urls": ["string"],
8 "mask_url": "string",
9 "aspect_ratio": "string",
10 "resolution": "string",
11 "background": "string",
12 "output_format": "string",
13 "enhance_prompt": true,
14 "seed": 0
15 }
16}model必填stringThe model name to use for generation.
"qwen/qwen-image-2-1"
callBackUrl可选stringCallback URL for task completion notifications. If omitted, no callback will be sent.
"https://your-domain.com/api/callback"
channel可选stringOptional APIPASS provider routing channel. The default channel is auto; APIPASS automatically allocates tasks across available providers based on real-time pricing and stability metrics. starter is ultra-low-cost with limited quotas and weaker stability. regular is standard and much cheaper than official APIs with moderate stability. official uses the model native API with high stability, fast task execution, and official pricing.
The default channel is auto. APIPASS automatically allocates tasks across available providers based on real-time pricing and stability metrics. starter is ultra-low-cost with limited quotas and weaker stability; regular is standard and much cheaper than official APIs with moderate stability; official uses the model native API with high stability, fast task execution, and official pricing.
可用选项:
auto
The input object contains the following parameters:
input.prompt必填stringText prompt for text-to-image generation or image editing. Maximum length: 5000 characters.
Describe the subject, style, lighting, composition, and editing target. The prompt can be used for both text-to-image generation and image editing.
"A cinematic mountain lake at sunrise with mist and reflective water"
input.image_urls可选array(URL)Reference image URLs. Supports 1 to 10 images. Supported formats: JPEG, PNG, and WebP. Maximum size: 30 MB and 25 MP per image.
When this list is non-empty, the adapter automatically uses the upstream image-to-image route. Compatible aliases are accepted for legacy integrations, but new calls should use image_urls.
["https://example.com/reference.png"]
input.mask_url可选stringOptional mask URL for local image editing. White areas indicate the generation area.
Use a white mask region to indicate the area to regenerate. This parameter is only valid with exactly one reference image and cannot be used with transparent background.
"https://example.com/mask.png"
input.aspect_ratio可选stringAspect ratio of the generated image.
For text-to-image requests, the default is 3:2. For image-to-image requests, the default is auto. auto only applies when a reference image is provided, and this parameter is ignored during mask-based local editing.
可用选项:
"16:9"
input.resolution可选stringResolution of the generated image. Default: 1K.
2K generation may take longer than 1K.
可用选项:
"1K"
input.background可选stringBackground mode for the output image. Default: opaque.
Transparent background outputs an image with alpha in PNG/WebP format and cannot be combined with jpeg or jpg output.
可用选项:
"opaque"
input.output_format可选stringOutput image format. Default: png.
jpg is accepted and normalized to jpeg. jpeg cannot be combined with transparent background.
可用选项:
"png"
input.enhance_prompt可选booleanWhether to rewrite the prompt into a more complete scene description before generation. Default: true.
true
input.seed可选numberRandom seed used to reproduce the same result when the model version, prompt, and parameters are unchanged. Default: 0.
0
1curl -X POST "https://api.apipass.dev/api/v1/jobs/createTask" \
2 -H "Content-Type: application/json" \
3 -H "Authorization: Bearer YOUR_API_KEY" \
4 -d '{
5 "model": "qwen/qwen-image-2-1",
6 "callBackUrl": "https://your-domain.com/api/callback",
7 "channel": "auto",
8 "input": {
9 "prompt": "A cinematic mountain lake at sunrise with mist and reflective water",
10 "aspect_ratio": "16:9",
11 "resolution": "1K",
12 "background": "opaque",
13 "output_format": "png",
14 "enhance_prompt": true,
15 "seed": 0
16 }
17 }'1{
2 "code": 200,
3 "message": "success",
4 "data": {
5 "taskId": "task_12345678"
6 }
7}codeStatus code, 200 for success, others for failure
messageResponse message, error description when failed
data.taskIdTask ID for querying task status and results

flux
flux
Generate photorealistic images with robust text
起价
0 积分
Nano Banana 2 Lite, aka Gemini 3.1 Flash-Lite Image, is a newly released image model from Google DeepMind built for fast 1K image generation, efficient creative iteration, prompt-based image editing, and responsive visual output in product-facing creative scenarios.
起价
0 积分
OpenAI
OpenAI
Generate intentionally designed visuals with less prompting using the latest OpenAI GPT Image 2 model.
起价
0 积分
Nano Banana now allows you to freely draw on edited images, bringing you a completely new experience. Simply click on the image to draw whatever you want. You can also add text and details, and Gemini will do the rest.
起价
0 积分
Fast image generation model built on Gemini 3.1 Flash Image. Supports text-to-image, image editing, multi-image fusion with up to 14 reference images, Google Search grounding, and outputs up to 4K resolution.
起价
0 积分
seedream
seedream
起价
0 积分
grok
grok
Grok Imagine is xAI’s AI video generation model for creating high-quality videos with realistic motion, native audio, and synchronized speech. Grok Imagine 1.5 introduces faster generation speeds, improved motion physics, and stronger visual consistency for modern AI video applications.
起价
0 积分
qwen
qwen
Generate realistic 2K images with structured text
起价
0 积分
seedream
seedream
起价
0 积分
OpenAI
OpenAI
Bring your ideas to life with sharper, faster, more precise image generation using the GPT Image 2.5 API.
起价
0 积分
wan
wan
Generate images with cinematic precision through high-fidelity text-to-image synthesis and structural editing.
起价
0 积分
seedream
seedream
Gemini said Generate images with lightning-fast efficiency and high-fidelity detail through a streamlined, lightweight architecture optimized for instant creative expression.
起价
0 积分
qwen
qwen
Generate realistic 2K images with structured text
起价
0 积分
Luma
Luma
起价
0 积分
seedream
seedream
起价
0 积分