Wan 2.7

• 🎬 Text-to-video generation from Alibaba's Wan family
• 🎯 Photorealistic output with strong subject coherence
• 🏢 Built by Alibaba, China's cloud and AI giant
• 🖼️ Consistent characters and motion across frames
• ⚡ Newest generation of the Wan video line

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 generative video family, turning written prompts into photorealistic clips that hold subject identity and motion steady across frames. Released in April 2026, it is the newest generation of the Wan text-to-video line, succeeding Wan 2.6 from late 2025 and the earlier Wan 2.5 Preview and Wan 2.2 builds.

The 2.7 release spans a coordinated set of modalities: alongside this text-to-video model sit an image-to-video variant, a reference-to-video model for guided generation, a video editing option, and text-to-image and Pro image generators. It complements Alibaba's broader catalogue, which includes the Qwen language and vision models and the HappyHorse video line.

Wan 2.7 is best suited to creators who need clean, coherent footage from a text description alone — product shots, narrative scenes, and stylized visuals where character consistency and lifelike rendering matter. Its emphasis on subject coherence makes it a strong pick when frame-to-frame stability is the priority.