Wan 3.0 by Alibaba online: video with sound from text and photos up to 1080p and 15 seconds. Versions 3.0, 3.0 Prime and 2.7.

Wan is Alibaba's video model family (Tongyi Wanxiang). Version 3.0 generates sound out of the box and renders up to 1080p, Prime is the highest quality, and 2.7 is the previous generation at a lower price.
All versions work from text or photos: one photo becomes the first frame, two become the first and last.
Yes, 3.0 and 3.0 Prime generate sound automatically. Version 2.7 makes silent video.
It depends on version, duration and resolution — e.g. 5 seconds of Wan 3.0 at 480p costs 4 tokens. The exact price is shown before you start.
Start generating with Wan Video for free. Free tokens on signup.