Alibaba has launched Wan2.6, an updated series of AI visual generation models, on December 16 in Hangzhou.

Wan2.6 allows users to appear in AI-generated videos using their own likeness and voice, and supports multi-shot storytelling with extended durations.

The series features a new reference-to-video generation model, Wan2.6-R2V, which lets users upload videos of people, animals, or objects, and generate new scenes that preserve the original appearance and voice.

Alibaba said Wan2.6-R2V is the first model in China to enable this type of consistent visual and audio insertion for video creation.

🔗 Source: Alibaba

Recent Alibaba developments