Vidu Q4 Preview AI Video Generation

Vidu Q4 Preview is the newest video model from ShengShu Technology, the Beijing lab behind Vidu. FlyAIgh added it in October 2026 while ShengShu still labels it a preview, so its output may change as the model is updated. It does two things. Image-to-video takes one still as the first frame and keeps that image's framing and look. Reference-to-video takes 1 to 12 images — characters, places, props — and composes them into one new shot in 16:9, 9:16, 1:1, 4:3 or 3:4.

Clips run any whole number of seconds from 3 to 16 at 540p, 720p, 1080p, 2K or 4K. Dialogue and sound effects are generated with the picture; sound is a switch, and it does not change the price. In reference mode you can also attach up to 3 MP3 tracks as audio references, and prompts can run to 5,000 characters.

What it does not do: there is no text-to-video and no first-and-last-frame mode. Vidu Q3 Pro covers both and has its own page.

Q4 costs 6 credits per second at 720p — 30 credits for a 5-second clip and 96 for the full 16 seconds. For a 5-second clip, 540p is 10 credits, 1080p is 37, 2K is 60 and 4K is 120; a 16-second 4K clip is 384. Reference images and audio references are not charged, so twelve references cost the same as one. The 540p tier is a cheap way to test a prompt or a reference set before paying for the finished resolution. The exact total is shown before you generate.

Pick Q4 when you are working from images — a single key visual to animate, or a set of references that must appear together in one shot — and when you want a higher resolution or a longer clip than Q3 Pro's fixed 4, 5 or 8 seconds at up to 1080p. Pick Q3 Pro when the shot starts from text alone or has to end on a frame you supply. Like every model on FlyAIgh, Vidu Q4 runs on the same credit wallet, with no separate Vidu account.