Wan 2.7

• 🆕 Alibaba's latest Wan text-to-video model, generating clips from prompts.
• 🔧 Adds first-and-last-frame and video-continuation controls per Alibaba Cloud docs.
• 📏 Supports 720p and 1080p output options.
• 🎯 Alibaba describes photorealistic generation with strong subject coherence.^provider^
• 🌐 Released early April 2026; API via Alibaba Cloud Model Studio.
• 🏢 Part of a coordinated Wan 2.7 release wave.
• 👁️ Companion image-to-video, reference, and edit modes available.

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 text-to-video member of Alibaba's Wan family, generating video directly from a text prompt. It ships alongside companion modes in the same release wave, including Wan 2.7 for image-to-video, Wan 2.7 Reference, and Wan 2.7 Edit, plus the image generators Wan 2.7 and Wan 2.7 Pro.

Compared with its predecessor Wan 2.6 (December 2025), Wan 2.7 continues the family's text-to-video line within Alibaba Cloud's Model Studio. Per Alibaba Cloud's own documentation, the Wan generation API supports features such as first-frame-to-video, first-and-last-frame, and video continuation, with 720p and 1080p resolution options.

Alibaba describes the Wan family as producing photorealistic generation with strong subject coherence across frames, and the documented controls give creators frame-level guidance over how a clip starts and ends.

The model targets creators needing controllable short-form video — marketing clips, social content, and reference-driven character work — and is accessible through Alibaba Cloud's Model Studio (DashScope) APIs. Specific licensing terms vary by version, so consult the official model card before commercial use.