Table of Contents
Alibaba has officially rolled out Alibaba Wan3.0, the latest version of its AI video generation model, with the ability to create videos of up to 30 seconds from a wide range of inputs, including documents, spreadsheets, presentations and web pages.
The launch came on August 24, one day after Alibaba announced an HK$80 billion ($10.2 billion) share placement in Hong Kong. The company said 100% of the net proceeds would be used to strengthen its full-stack AI capabilities, including AI infrastructure, chips and model development.
What Is Alibaba Wan3.0?
Alibaba Wan3.0 is designed to expand AI video creation beyond conventional text-to-video workflows.
The model can accept:
- Text
- Images
- Audio
- Video
- Documents
- Spreadsheets
- Presentations
- PDFs
- Web pages
This allows users to provide content such as a presentation or PDF and generate a video based on the information contained within it.
Wan3.0 can produce videos of up to 30 seconds in a single generation, twice the maximum duration offered by its Wan2.7 predecessor. It supports resolutions of up to 1080p.
Alibaba says the model has also been improved in areas including instruction following, consistency between shots and audio quality. However, these performance claims are company-provided and should not be treated as independently verified benchmarks.
From Documents to Video
One of the most notable features of Alibaba Wan3.0 is its ability to work with information-heavy files.
Instead of manually converting a presentation or document into a video script, users can provide supported files as inputs. This could be useful for corporate presentations, product information, marketing materials, training content and other applications where existing written material needs to be converted into video.
Alibaba has also highlighted applications including short dramas, advertising, tourism promotion and other creative uses.
The feature does not mean every document will automatically become a finished production-quality video. The resulting output can depend on the source material, instructions and generation process.
Wan3.0 Pricing
Alibaba has introduced usage-based pricing based on output resolution.
| Resolution | Price per second | Cost for 30 seconds | Approx. cost per minute |
| 480p | $0.05 | $1.50 | $3 |
| 720p | $0.10 | $3 | $6 |
| 1080p | $0.20 | $6 | $12 |
A temporary 30% discount on API pricing is available from August 24 through September 23 on selected platforms, according to reporting on the launch.
For comparison, Google’s Veo 3.1 standard tier has been reported at $0.40 per second, although pricing comparisons alone do not establish which model delivers better overall quality.
Availability of Alibaba Wan3.0
Although Alibaba officially expanded the model’s availability on August 24, Alibaba Wan3.0 had already entered public beta on August 6 through Alibaba Cloud’s Model Studio and Qwen Cloud.
Alibaba is also preparing a consumer-facing version through wan.video, while broader API access is being rolled out.
Unlike some previous Wan releases, Wan3.0’s model weights are closed, marking a shift from Alibaba’s earlier open-weight approach to its video models.
The Launch Comes Amid Heavy AI Investment
The timing of the Wan3.0 rollout is significant because it came immediately after Alibaba’s major Hong Kong share placement.
The company announced the HK$80 billion fundraising, with all net proceeds earmarked for its full-stack AI capabilities. Reuters reported that the planned investment includes chips, infrastructure and the development and deployment of AI models.
Alibaba’s AI push has also increased its expenses. Recent reporting showed a 75% year-over-year decline in quarterly net profit, while the company continues to invest heavily in AI infrastructure.
The financial developments provide broader context for Alibaba’s growing focus on commercialising AI products such as Wan3.0.
What Wan3.0 Means for AI Video Generation
Alibaba Wan3.0 demonstrates how AI video tools are increasingly moving beyond simple text prompts toward multimodal workflows.
Its ability to process documents, presentations, spreadsheets and web pages gives businesses another way to turn existing information into visual content. At the same time, the model’s closed weights and limited independent benchmarking mean its long-term position against competing video-generation systems remains uncertain.
For now, Wan3.0 represents Alibaba’s latest attempt to combine multimodal AI, video generation and commercial cloud services in a single offering.
FAQs
What is Alibaba Wan3.0?
Alibaba Wan3.0 is Alibaba’s latest AI video generation model, capable of creating videos up to 30 seconds long from multiple forms of input.
Can Wan3.0 generate videos from PDFs and presentations?
Yes. Alibaba says the model supports documents such as PDFs, spreadsheets and presentations, along with web pages and other multimedia inputs.
How much does Wan3.0 cost?
Pricing ranges from $0.05 per second at 480p to $0.20 per second at 1080p. A 30-second 1080p video therefore costs $6 at the listed rate before discounts.
Is Wan3.0 open source?
No. Unlike some earlier Wan models, Wan3.0’s model weights are closed.
When did Wan3.0 become available?
The model entered public beta on August 6, while Alibaba expanded its rollout on August 24, 2026.
Conclusion
Alibaba Wan3.0 expands AI video generation by allowing users to create videos not only from prompts and media but also from information-rich sources such as PDFs, spreadsheets, presentations and web pages.
Its 30-second generation limit, 1080p support and usage-based pricing make it a notable addition to the growing AI video market. However, independent benchmarking remains limited, so claims about its overall quality should be treated cautiously.
The launch also arrives during Alibaba’s much larger AI investment push, making Wan3.0 both a new creative AI product and part of the company’s broader strategy to build and commercialise its full-stack AI capabilities.