Wan 2.7 is an AI video model released in the beginning of April 2026 that makes short video clips from text or images. It fits into the video generation category and supports both text to video, image to video and video to video tasks.
The model lets you edit or recreate videos using text prompts, reference images or style changes. You can guide how the final video looks just by describing it or giving a sample.
It also supports reference to video features. It can match a person’s look and even their voice tone, using up to five real people at once. And it can handle scenes with multiple characters interacting, keeping them consistent across the clip.
Wan 2.7 lets you pick video length in small steps from 2 to 15 seconds. You can choose 2, 3, 4 and so on up to 15… so you get simple control over clip duration.
The resolution is 720p or 1020p.
The model focuses on better motion and smoother video flow. It tries to keep things stable from frame to frame, so objects don’t shift or flicker much. It also improves how camera moves look, like pans or zooms, making clips feel more natural.
Earlier versions already did solid image to video work, and this version builds on that with more stable motion and better response to prompts.
If you'd like to access this model, you can explore the following possibilities:
