Wan 2.1 Pro

• 🎬 Alibaba Tongyi Lab's image-to-video model from the Wan family
• 🖼️ Animates a single still image plus text prompt
• 👁️ Emphasizes photorealism and subject coherence across frames
• 🏢 Built by Alibaba's Tongyi Lab
• 📏 Accepts first-frame image input via Model Studio API
• ⚡ Image-to-video jobs typically run via asynchronous API polling
• 🌐 Part of a broad Wan lineup spanning text, image, video

Alibaba

Alibaba Group is a Chinese multinational technology company founded in 1999 and headquartered in Hangzhou, Zhejiang. Originally built around e-commerce and cloud computing, Alibaba has become one of the most prolific contributors to open-weight AI research,…

Wan 2.1 Pro is the image-to-video member of Alibaba's Wan video family, developed by the company's Tongyi Lab. It takes a single still image together with a text prompt and animates it into a short clip, with the catalog highlighting photorealistic generation and strong subject coherence across frames. Image-to-video tasks are generated asynchronously, typically completing in a few minutes per request.

Within the wider lineup, Wan 2.1 sits at the foundation of a fast-moving generation cycle. Later siblings such as Wan 2.6 and Wan 2.7 extended duration and resolution, with documentation describing image-to-video output up to 1080p and as long as 15 seconds in the 2.6 generation. The Wan family on Model Studio also supports first-and-last-frame conditioning, where a start and end image guide the generated motion.

For users choosing within the family, Wan 2.1 Pro represents the established, photorealism-focused image-to-video option, while the Wan 2.5 Preview and subsequent releases carry the architectural and feature upgrades. Always validate current capabilities against Alibaba's own model documentation before production use.