Vidu

Q3 is the premium tier and Q3 Turbo is the faster, cheaper one with the same feature set, including native audio. Turbo costs roughly half as much, which makes it right for drafts and idea testing, while Q3 is the better choice for narrative fidelity in finals. Both support clips from 1 to 16 seconds and resolutions from 360p to 1080p.

Use Q2 for tight 4–8 second character shots with expressive faces and stable camera moves; use Q3 when you need native audio or clips longer than 8 seconds. Q2 offers image-to-video Turbo for fast drafts and Pro for final-quality detail. Q2 text-to-video has no end-frame input, and its turbo_mode only applies to image-to-video.

Q3 produces 1 to 16 second clips, but 4–5 seconds is the most stable range. Drift appears past roughly 8 seconds, so don't expect a single steady shot at 12–16 seconds. Q2 tops out at 8 seconds.

Yes, on Q3 and Q3 Turbo, with the Generate Audio toggle on. Dialogue, effects and music come out in one pass. Put spoken lines in quotes and name the ambience you want, such as rain or footsteps. English, Japanese and Chinese are the officially supported dialogue languages. Q2 offers only background music, and only at 4 seconds.

On Q1 text-only prompts, set Style to Anime. On other models and for image-driven runs the style toggle has no effect, so describe cel shading or 2D illustration in your prompt or start from an illustrated first frame, which animates especially cleanly.

It trails the best models on photoreal skin close-ups, complex physics, extreme action and readable on-screen text. Stacking several actions in one prompt also hurts results. It works best with one dominant motion and one camera move, on a character-led shot.