By CoinAINews Staff |

Alibaba has officially launched Wan 3.0, its latest AI video generation model — and it can turn your PowerPoints and PDFs into 30-second videos.

The model can generate up to 30 seconds of video in a single generation, with output reaching 1080p resolution. But the longer video length is only part of the story. Wan 3.0 is also built to work with different types of reference material, including documents, presentations, images, audio and video.

That puts Alibaba's latest model in an interesting position. It is not being pitched only as a tool for making cinematic clips from text prompts. It is also aimed at people who already have information sitting inside reports, presentations and other business documents and want to turn that material into video.

And at $6 for a 30-second 1080p generation, the pricing is likely to attract plenty of attention.

What Wan 3.0 Can Actually Do

30-Second Video Generation

One of the clearest upgrades in Wan 3.0 is its ability to generate up to 30 seconds of video in a single generation.

That gives creators more space to build a complete sequence without automatically breaking an idea into several shorter clips.

Alibaba says Wan 3.0 supports native 30-second generation and can produce output at 480p, 720p and 1080p resolutions.

The longer generation window could be useful for advertising, short-form storytelling, product demonstrations and other content where a few seconds simply isn't enough to communicate the full idea.

But there is an important distinction here: a longer generation does not automatically mean a better video. Consistency, movement, scene transitions and instruction following still determine whether the final clip is actually useful.

PDFs and PPTs Are Where It Gets Interesting

This is probably the feature that will get businesses paying the closest attention.

Wan 3.0 can use documents, presentations, spreadsheets and other structured information as part of its video-generation workflow.

Businesses already produce enormous amounts of content in the form of presentations, reports, spreadsheets, training materials and marketing documents. Wan 3.0 is designed to turn that static information into video.

Imagine a company preparing a 20-slide product presentation.

Normally, converting that presentation into a video could mean writing a separate script, selecting visuals, recording narration and sending everything through an editing workflow.

A document-aware AI system can potentially reduce some of those steps by using the original material as the foundation for the video.

That doesn't mean you can upload any random PDF and instantly get a finished Hollywood-style production. Human review will still matter, especially when the source material contains complicated information, branding requirements or important factual details.

Still, the workflow is significant.

The AI isn't simply generating something from a blank prompt. It is using information that already exists.

Multiple Reference Inputs

Wan 3.0 also takes a broader multimodal approach to video creation.

Users can provide different reference assets to help guide the generation process. These can include images, videos and audio alongside other forms of input.

That matters when a creator needs more control over what appears in the final video.

For example, instead of describing a character entirely through text, a reference image can help establish the character's appearance. Additional visual or audio references can then be used to guide other parts of the scene.

The goal is to make the model less dependent on pure text prompting and more useful for workflows where creators already have source material.

Editing and Video Extension

Wan 3.0 is not limited to creating a new video from scratch.

Alibaba also positions the model for video editing and extension, allowing users to work with existing footage and make changes through instructions.

That can be useful for creators who don't want to regenerate an entire clip just because one part needs to change.

Video generation is increasingly moving in this direction. Instead of treating AI as a one-click generator, companies are trying to turn it into a complete production assistant that can generate, modify and extend content.

Wan 3.0 is part of that shift.

How Much Does Wan 3.0 Cost?

This is where Alibaba's model gets particularly interesting.

Alibaba Cloud's current pricing lists Wan 3.0 at different rates depending on resolution:

At the highest listed resolution, a 30-second 1080p video costs approximately $6 based on the published per-second rate.

That doesn't necessarily make Wan 3.0 the cheapest option in every possible workflow, because real production costs can also depend on retries, editing, additional assets and platform usage.

Still, the headline number is straightforward: 30 seconds of 1080p generation for about $6.

For businesses producing large amounts of short-form content, that pricing could become an important part of the decision.

Alibaba's AI Push Comes at the Same Time

The Wan 3.0 launch comes shortly after Alibaba announced an HK$80 billion, or roughly $10.2 billion, Hong Kong share placement.

Reuters reported that Alibaba plans to use the proceeds to support its broader AI development, including investment in AI infrastructure, chips and models.

The timing is notable because Alibaba has been aggressively expanding its AI ambitions while facing the enormous infrastructure costs that come with running increasingly capable models.

Wan 3.0 therefore isn't an isolated product launch.

It fits into Alibaba's wider strategy of building an AI ecosystem that covers models, cloud computing and applications.

The company already has its Qwen family of large language models, while Wan gives it a stronger position in the increasingly competitive AI video market.

Why Document-to-Video Could Matter

The biggest opportunity may not be filmmakers at all.

Businesses already produce enormous amounts of content in the form of presentations, reports, spreadsheets, training materials and marketing documents. Wan 3.0 is designed to turn that static information into video.

That could open up a much larger market than traditional AI video generation.

Think about internal training.

A company might have a lengthy employee handbook or a presentation explaining a new process. Turning that information into a short video could make it easier for employees to consume.

The same idea works for marketing teams.

A product presentation could potentially become a promotional video. A travel company could turn destination information into visual content. An education provider could transform learning material into short explanatory clips.

The possibilities are broad because the source material already exists.

The challenge is making sure the AI understands that material correctly.

A visually impressive video isn't much use if it changes numbers, misunderstands a product specification or leaves out an important part of the original information.

That's why human review is likely to remain an important part of the workflow.

Wan 3.0 vs. the Competition

Alibaba isn't entering an empty market.

Google, OpenAI and several specialist AI companies are developing increasingly sophisticated video-generation systems. The technology is improving quickly, and new models can make comparisons outdated within months — sometimes weeks.

Wan 3.0's strongest selling points appear to be its combination of longer generation, multimodal inputs, document-based workflows and relatively low listed pricing.

But price isn't everything.

Users will ultimately care about the quality of the output, consistency between scenes, how accurately the model follows instructions and how much editing is required before a video is ready for publication.

Independent testing will be important here.

Wan 3.0 is still relatively new, so broad third-party comparisons with competing models remain limited. Alibaba's own performance and capability claims should therefore be viewed as company-reported until more independent evaluations become available.

Availability

Wan 3.0 is available through Alibaba's AI platforms, including Alibaba Cloud Model Studio and Qwen Cloud, with access and individual features depending on the service and region.

Alibaba has also been expanding the model's availability for different creative and commercial use cases.

That could eventually put the technology in front of a much wider audience than developers experimenting with AI video generation.

For businesses, the more interesting question may be whether Wan 3.0 can fit into existing workflows rather than simply whether it can generate impressive clips.

What Comes Next

AI video generation is gradually moving away from the simple idea of entering a sentence and receiving a video.

The next stage is likely to involve models that understand documents, images, presentations, audio, existing footage and other business information and then use all of it to build content.

Wan 3.0 is clearly moving in that direction.

The model's document capabilities could prove especially useful if Alibaba can make the process reliable enough for professional use.

But reliability is the part that will take time to judge.

A business needs more than an impressive demo. It needs consistent results, accurate information, predictable costs and enough control to make the final output match its requirements.

The Bottom Line

Alibaba's Wan 3.0 is more than a straightforward upgrade to an AI video generator.

It can generate up to 30 seconds of video, supports output up to 1080p, works with multiple forms of reference material and is designed to bring documents and presentations into the video-generation process.

The pricing is another major attraction. At the listed $0.20 per second for 1080p, a 30-second video works out to roughly $6 before any additional usage or production costs.

Whether Wan 3.0 can match the very best competing models on overall quality remains an open question. Independent testing and wider real-world use will tell that story.

But Alibaba may be aiming at something bigger than another AI video demo.

Wan 3.0 could be more about changing how businesses turn information into content — and less about generating another impressive AI demo.

---

Disclaimer: This article is for informational purposes only. AI models, pricing and availability can change, so users should verify current specifications and costs before relying on them.