Skip to content

Video

Vidu Q3 Turbo

Video

Available on Ofrai — chat, images, and video in one place with shared credits.

115 creditsModel details

Video studio

Preview

No clip yet

Your latest clip preview will show here after generation.

Open studio

Example prompts

Choose an example to load its prompt and recommended settings into the Vidu Q3 Turbo creator above.

How it works

Create with Vidu Q3 Turbo in three steps

  1. 01

    Choose your inputs

    Keep Vidu Q3 Turbo selected, then add a prompt and any supported reference assets.

  2. 02

    Set the output

    Choose the aspect ratio, resolution, and available Video options before generating.

  3. 03

    Generate and refine

    Review the result, adjust the prompt or settings, and generate another version from the same place.

Vidu Q3 Turbo is the tier the documentation recommends on arithmetic. The reference calls it the fast model, notes that it supports smart scene switching, and describes it as the fastest to generate with the best cost-performance ratio in the line. Pro is about twice its price, so the question this page has to answer is narrow: is there anything on a given job that justifies doubling the budget? For most iteration, no — because Turbo and Pro expose the same parameter surface.

Same 1 to 16 second duration, same three resolutions, same zero/one/two reference-image semantics, same audio default. Turbo is where shots get moved, and it is where a job with forty renders stops being a cost argument.

The reference puts the value verdict on this tier

Two phrases in the official documentation describe Turbo, and both are comparative: it generates fastest, and it has the best cost-performance ratio. That is a throughput claim, not a fidelity claim, and it is the whole reason to choose it. Think about where generation time actually lands in a video job. Prompt drafting is fast. Reviewing a clip is slow.

Reading four variants side by side to decide which direction to take is the part where a fast model changes behaviour rather than just saving time: a turnaround measured in a comfortable wait becomes a turnaround you can iterate inside, and teams that iterate inside a session ship more finished shots than teams that batch overnight. The one documented capability Turbo carries on its own is smart scene switching — the ability to move between locations within a generated clip without the seam collapsing.

Turbo and Pro share an identical parameter surface

This is the part that makes the decision easy and is frequently missed. Duration on Turbo is 1 to 16 seconds, default 5, exactly as on Pro — so the base model's 3-second floor does not apply here and short inserts are available at the cheap tier. Resolution is the same three options, 540p, 720p and 1080p, with 720p the default.

Reference images behave identically: none for text to video, one for a first frame, two for a first and last frame pair, with the larger multi-reference path reserved for the base model. Audio defaults to true on Turbo too, so generated speech and ambience are available without paying for the premium tier. So you are not trading away access to features by choosing Turbo on those parameters.

The capability the documentation states for Pro and not for Turbo is the audio-video sync and storyboard pairing, and that is what the higher price is actually buying.

One-second renders are the real innovation here

A one-second clip is almost useless as a deliverable and extremely useful as a unit of work. It answers questions that a five-second render cannot answer cheaply: does this camera move work at all, does the first frame I supplied hold, is the framing right before I spend a minute of wall clock finding out.

Because Turbo accepts a one-second duration at the cheapest tier in the family, the shortest possible test is also the least expensive test, and the loop between writing a prompt change and seeing the result is dominated by generation rather than by queueing. Set a habit here that pays off later: keep the subject, action and style wording identical between passes and change one thing — the camera, the framing, the moment in the action — so that when a render improves you know which edit did it.

Otherwise you have motion instead of information.

Scene switching, shot switching and storyboards are three different things

The Q3 family uses three distinct words for camera behaviour, and choosing between models by vibe rather than by these words is how people end up paying for the wrong one. Turbo is described with smart scene switching: the model moves the action between locations or settings within a clip, which is what you want when the beat involves a change of place. The base model is described with smart shot switching, alongside the best-in-family multi-camera consistency, which is about holding one subject recognisable across different angles.

Pro is described with storyboard video generation, which is about laying out shots. If your requirement is a change of location, that is Turbo's word. If it is one identity surviving several angles, that is the base model's. If it is planning a sequence, read Pro's storyboard description before assuming Pro is your default.

540p is a legitimate draft tier, not a compromise

Three resolutions, one default, and a price difference within the model family — which means the draft settings are where most of the savings are actually taken. 540p is enough to judge composition, lighting direction, colour temperature and whether the action reads at all. It is not enough to judge texture, fine detail or whether a face holds up, and that distinction should decide what you look at. Review at 540p and 720p, render the keeper at 1080p.

Aspect ratio takes 16:9, 9:16 and 1:1, with the family noted as supporting arbitrary ratios, and passing a reference image often lets the image decide the ratio — another reason to set the ratio explicitly when the deliverable has a fixed frame. The common failure is drafting at 1080p because it is available, which pays the premium tier's price for frames nobody watched closely.

Asynchronous jobs are what make throughput possible

Submission returns a task identifier and the result is collected later, which is the mechanism underneath every throughput claim. Because a job does not block you, the practical workflow is to fan several variants out and collect them as they land: one Turbo job for the framing you are unsure about, one for the alternate camera height, one for the version with the reference image instead of pure text. Each variant is a cheap, independent question, and the identifier is what keeps them from getting confused.

Keep a small table of job ID, prompt and draft settings, because the fastest way to lose a good result is to be unable to reproduce the request that produced it. Also keep the failures: a rejected or odd render is still evidence about the prompt, and re-rolling without recording what changed throws that evidence away.

The one-sentence rule for choosing

Draft and explore on Turbo, and escalate to Pro only at the moment a shot stops changing and has to survive review at full fidelity. Escalation is cheap to reason about because the gap is a stated multiple rather than a quality mystery, and it is easy to check whether the jump was worth it: render the same shot on both tiers and look at what actually changed. If nothing changed, you were drafting at the wrong price and Turbo was the correct answer all along.

If the difference is visible, it now has a budget owner. Two exceptions to the rule: if the deliverable needs generated sound tightly matched to the picture, the sync capability is documented on Pro, so that is a legitimate escalation reason rather than an aesthetic preference. And if the shot needs several reference images reconciled into one identity, Turbo's two-image mode is not enough and the base model is the right target.

FAQ

Why would I use Turbo instead of Pro if the parameters are identical?
Because the reference prices Pro at roughly two times Turbo and describes Turbo as the fastest with the best cost-performance ratio. Since duration range, resolution set, reference-image rules and the audio default are the same on both, choosing Turbo is a budget decision rather than a capability sacrifice — the documented exception is Pro's audio-video sync and storyboard support.
Can Turbo render clips shorter than 3 seconds?
Yes. Turbo, like Pro, accepts 1 to 16 seconds and defaults to 5. The 3-second floor applies to the base `viduq3` model only, so very short test renders are available at the cheap tier.
What is smart scene switching, and how is it different from the base model's smart shot switching?
Smart scene switching is the behaviour documented for Turbo and concerns moving between locations or settings within a clip. Smart shot switching is documented for the base model and concerns holding one subject recognisable across different camera angles, alongside its multi-camera consistency and multi-reference support.
How does the asynchronous job flow affect my workflow?
Submission returns a task identifier and the result is collected later, so variants can run in parallel instead of one at a time. Record the job ID, prompt and draft settings together, and keep the failed attempts as evidence about what the prompt is doing.

FAQ

Can I use Vidu Q3 Turbo directly on this page?
Yes. The creator above is the working Ofrai studio with Vidu Q3 Turbo selected by default.
Where are my finished results saved?
Signed-in generations are kept in your Ofrai library so you can download or revisit them.
Which prompts work best with Vidu Q3 Turbo?
Describe the subject, setting, visual style, lighting, and camera or motion clearly. The examples above provide practical starting points.
Can I edit the example prompts before generating?
Yes. Load any example, then change its wording, aspect ratio, references, or other supported settings in the creator.

Compare other Video models that can handle similar creative work to Vidu Q3 Turbo.

Commercial use depends on the selected model and model terms.