← All models

Wan2.5-I2I-Preview

Published Publisher Alibabawan2.5-i2i-preview

Overview

Wan2.5-I2I-Preview is a preview image-editing model in Alibaba’s Wan series, served through Alibaba Cloud Model Studio. It uses text instructions and reference images to produce edited images. Alibaba describes improved instruction following, consistent multi-image reference generation, and strong text generation within images. Suitable uses include editing an existing image and combining elements from multiple images. (alibabacloud.com)

Release date

Alibaba’s release record lists the model under September 23, 2025 for Beijing and September 24, 2025 for the Singapore international deployment. Official release record. (alibabacloud.com)

Capabilities and limits

  • Supports single-image editing and multi-image fusion; it is not a standalone text-to-image model. (help.aliyun.com)
  • The API accepts up to three reference images, each up to 10 MB, with width and height between 384 and 5,000 pixels. Supported inputs include JPEG, JPG, PNG without an alpha channel, BMP and WEBP. (help.aliyun.com)
  • Output is PNG. The documented output pixel count ranges from 768 × 768 to 1280 × 1280, with an aspect ratio from 1:4 to 4:1; the documented default is 1280 × 1280 pixels, with the approximate aspect ratio derived from the input. The model listing specifies up to four outputs. API reference. (help.aliyun.com)
  • The model information page lists function calling, structured outputs, web search, context caching, batch inference and fine-tuning as unsupported. Context length is not specified. Model information. (alibabacloud.com)

Pricing

Checked November 18, 2026. The listed API prices charge for output images only; input is not separately billed. The official pricing tables state that this model has no free quota. These are platform- and region-specific prices, not currency conversions. (help.aliyun.com)