Chinese tech giant Alibaba has introduced Wan2.7-Video, a new AI-powered video generation model that promises unprecedented control over video content creation. Unlike previous tools focused solely on visual quality, Wan2.7-Video emphasizes flexibility and precision in video production. The model, currently available only for internal testing at wan.video, is not open-source, marking a strategic shift in Alibaba’s AI initiatives.
The new tool offers several standout features. Users can fix characters in videos by uploading up to five images and a voice sample, ensuring consistency across frames. This eliminates the need for manual editing in post-production. Additionally, the model allows real-time adjustments using simple text prompts, enabling creators to modify videos on the fly without starting from scratch. For narrative-driven projects, Wan2.7-Video includes a full production cycle, integrating object cloning, style transfers, and logical story extensions to maintain continuity.
Alibaba’s decision to keep Wan2.7-Video closed contrasts with its earlier open releases, such as version 2.2. Industry analysts suggest this may reflect the company’s focus on commercial applications rather than public accessibility. The move aligns with Alibaba’s broader push to dominate the AI and cloud computing sectors, particularly in video generation—a rapidly growing market.
The model’s capabilities position it as a potential game-changer for filmmakers, advertisers, and content creators. By combining generation, editing, and narrative tools into a single system, Wan2.7-Video could streamline workflows that traditionally require multiple software suites. However, its closed status limits immediate adoption outside Alibaba’s ecosystem.
For now, Wan2.7-Video remains a proprietary tool, with testing restricted to select users. Its release underscores Alibaba’s ambition to lead in AI-driven creative technologies, even as it prioritizes control over accessibility.
Resources: wan.video