Wan 2.7

• 🎥 Alibaba's Wan image-to-video model, released April 2026.
• 🖼️ Generates a video clip from a supplied first-frame image.
• 🌐 First-frame image-to-video via Alibaba Cloud Model Studio API.
• 🏢 Part of Alibaba's Wan photorealistic video family.
• 🎯 Family emphasizes subject coherence held across frames.
• 🆕 Newest image-to-video release, following Wan 2.6.

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 2.7 is the image-to-video member of Alibaba's Wan video family. Given a supplied first-frame image, it generates a video clip; Alibaba Cloud documents this first-frame image-to-video workflow through its Model Studio API. In the catalog's framing, the Wan family emphasizes photorealistic generation with subject coherence held across frames, making it suited to animating an existing still image into motion.

Compared with its same-family predecessor Wan 2.6, Wan 2.7 is the newer image-to-video generation in the family, which also includes the earlier Wan 2.5 Preview. Per the catalog, it released in April 2026, placing it after the December 2025 Wan 2.6 release.

Within the broader 2.7 lineup, this image-to-video model sits alongside siblings such as Wan 2.7 text-to-video and Wan 2.7 Reference, which target text-driven and reference-guided generation respectively. For workflows that start from an existing image asset, the first-frame image-to-video mode documented by Alibaba Cloud is the relevant entry point.