Wan

About Wan

What is Wan (2026 update)

Wan is Alibaba’s AI creative platform for generating and editing images and video from text prompts or reference images. It runs on Alibaba’s Tongyi Wanxiang model family and is aimed at creators, marketers, and small studios who want cinematic video generation without a full production pipeline.

What is new in 2026

Alibaba shipped several model upgrades through 2026, moving from Wan 2.5, which added native higher resolution output and one-pass audio and video synchronization, through Wan 2.7 in April 2026, which added native audio sync, first and last frame control, and multi-reference consistency, up to Wan 3.0 in August 2026. Wan 3.0 can generate up to 30 seconds of video in a single pass and accepts text, images, audio, video, and even documents like PPT, PDF, DOC, and XLS files as source input, letting you turn a slide deck or report directly into a video. On the API side, Wan 3.0 pricing runs $0.05 per output second at 480p, $0.10 at 720p, and $0.20 at 1080p, meaning a full 30-second 1080p clip costs about $6 through Alibaba Cloud Model Studio.

Key features

• Text-to-video and image-to-video generation with the latest Wan 3.0 model
• Up to 30 seconds of video in a single generation pass
• Accepts text, image, audio, video, and document inputs (PPT, PDF, DOC, XLS) as source material
• Native audio and video synchronization in one pass
• First and last frame control and multi-reference consistency for keeping subjects consistent across a clip
• Image generation, image editing, and image upscaling alongside video tools
• Watermark-free downloads and 1080p output on paid membership plans

Pricing in 2026

Wan’s consumer app at wan.video runs a free tier with a small watermark on exports, plus two paid membership tiers billed monthly or yearly, alongside one-time credit packs for occasional use.

Who should use it

Wan fits marketers, social content creators, and small agencies who want to turn a script, product photo, or even a slide deck into a short video clip without hiring a video editor, especially anyone comfortable working within a credit-based system. The API side, priced per second of output, suits developers building it into their own apps rather than using the consumer web app directly. It is less suited to teams needing long-form video beyond 30 seconds in one generation, or those with strict requirements around training data provenance, since Alibaba has not published full details on that for every model version.

Limitations and alternatives

Wan’s free tier adds a watermark to every export, and even paid tiers cap single generations around 30 seconds with Wan 3.0, so longer content still needs stitching multiple clips together. Pricing and model naming have changed several times in 2026 (2.5, 2.6, 2.7, then 3.0), so check the official site for the current default model before comparing costs to competitors. Alternatives worth comparing include Runway, which offers similar text and image-to-video generation with a longer track record for professional editing workflows, Kling AI, which is a comparable Chinese-built competitor with its own credit pricing, and Luma AI’s Dream Machine, which focuses on fast, shorter clip generation at a lower entry price.

Frequently asked questions

Is Wan AI free to use?

Wan has a free tier for basic image and video generation, but exports include a small watermark. Removing the watermark requires the Pro plan at $5/month or higher.

How long can a Wan generated video be?

As of Wan 3.0 in 2026, Wan can generate up to 30 seconds of video in a single pass, longer than earlier Wan 2.5 and 2.7 versions.

What inputs does Wan 3.0 accept?

Wan 3.0 accepts text, images, audio, video, and documents such as PPT, PDF, DOC, and XLS files as source material for generating a video.

How much does the Wan API cost?

Wan 3.0 API pricing on Alibaba Cloud Model Studio is $0.05 per output second at 480p, $0.10 at 720p, and $0.20 at 1080p, so a 30-second 1080p clip costs about $6.

Information last verified: September 2026.