视频 API已上线

wan-2-6 API

Wan 2.6 已通过 API in One 接入真实生产路由 /api/v1/videos/generations，后端节点为 node35。一个 API Key 即可调用并按成功请求计费。

获取 API Key 查看文档

真实能力

以下能力来自当前已接入的模型配置和后端路由，不展示未上线 API。

稳定生产路由

Backed by Alibaba's advanced research

统一接口

Up to 1080p resolution and 15s duration

透明计费

Text, image, and video input supported

快速接入

Unified billing with all other models

请求方式

浏览器端只调用同源 API，服务端 BFF 再转发到对应 Worker。

方法端点积分用途

POST/api/v1/videos/generations80 积分

Alibaba's Wan 2.6 video generation model. Supports text-to-video and image-to-video with high-fidelity output.

参数

promptstring

Text description of the video

必填

durationstring

Video duration: "5", "10", or "15" seconds

可选 · 默认 5

resolutionstring

Output resolution: "720p" or "1080p"

可选 · 默认 720p

image_urlsarray

Array of image URLs for image-to-video mode

可选

video_urlsarray

Array of video URLs for video-to-video mode

可选

示例请求

{
  "endpoint": "/api/v1/videos/generations",
  "headers": {
    "Authorization": "Bearer <API_KEY>",
    "Content-Type": "application/json"
  },
  "body": {
    "model": "wan-2-6",
    "prompt": "A time-lapse of cherry blossoms blooming in a Japanese garden, cinematic",
    "duration": "5",
    "resolution": "720p"
  }
}

任务状态查询

图像、视频、音频和工具任务创建后，使用真实任务端点查询结果。

GET/api/v1/tasks/{task_id}查询生成任务状态和结果。

价格

80–325 credits per video (~$0.80–$3.25)

基础价格$0.80

积分80

计费按成功请求

使用场景

这些场景可通过当前模型和对应后端节点实现。

Create marketing videos from text descriptions

Generate visual content for e-commerce product showcases

Build video generation features into SaaS platforms

Produce animated content from reference images

常见问题

How much does the Wan 2.6 API cost?

Credits depend on resolution and duration: 720p 5s costs 80 credits, 1080p 10s costs 220 credits, 1080p 15s costs 325 credits.

What resolutions does Wan 2.6 support?

Wan 2.6 supports 720p and 1080p resolution output for high-quality videos.

What durations are available?

Wan 2.6 supports 5, 10, and 15 second video generation.

Does Wan 2.6 support different input types?

Yes. Text-to-video, image-to-video (from reference images), and video-to-video (from reference video clips).