Skip to content

Video

Vidu Q3 Ad

Video

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

343 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 Ad creator above.

How it works

Create with Vidu Q3 Ad in three steps

  1. 01

    Choose your inputs

    Keep Vidu Q3 Ad 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.

Start with the honest part, because it changes how this page should be read. The Vidu Q3 generation documentation lists exactly three Q3 model identifiers: `viduq3-pro`, `viduq3-turbo` and `viduq3`. There is no ad-specific Q3 build listed there. This page therefore describes the commercial-placement brief — short, format-specific, product-faithful clips — and the Q3 family specification that brief is served from, and it names the documented model that owns each behaviour rather than claiming a fourth tier.

If you are here to compare benchmarks against a product-marketing engine, that comparison is not available from these docs, and no claim on this page should be read as one.

What the reference documents, in full and in the open

The complete list of Q3 models in the documentation is three. Pro is described as the high-quality model supporting audio-video sync and storyboard video generation, at roughly twice turbo's price. Turbo is described as the fast model supporting smart scene switching, fastest to generate and best on cost-performance. The base `viduq3` model is described as best in the line for multi-camera consistency, with smart shot switching and multi-reference image generation. Submission is asynchronous and returns a task identifier. That is the entire menu.

An 'Ad' variant is not in it, which means the useful question is not what an ad model does that others cannot — it is which of these three actually fits a placement, and this page answers that for the four things a placement genuinely needs: a faithful product, a defined frame, a short runtime and, if the buy includes it, sound.

The four things a placement needs, and the knob for each

A commercial slot imposes four constraints, and the Q3 family has a documented parameter for each. Product faithfulness is handled by reference images: one image becomes the first frame, so the thing on screen is the asset you supplied rather than something the model approximated from a text description. Frame is handled by the aspect ratio parameter, taking 16:9, 9:16 and 1:1, with the family noted as supporting arbitrary ratios beyond those.

Runtime is handled by the duration parameter, 1 to 16 seconds on Pro and turbo and 3 to 16 on the base model. Sound is handled by the audio boolean, which defaults to true across the family and generates speech and background effects from the prompt when enabled. Map those four to the brief before writing anything, because each one removes a class of revision later — and one of them, sound, costs nothing extra at the cheaper tier because it is on by default everywhere.

Keeping the product on frame: what one or two images do

The reference image parameter is where the difference between a usable placement and an unusable one lives. With no image supplied, the request is text to video and the model originates whatever it decides your product looks like, which is a gamble with the one asset that must not change. With one image, that image is the first frame: the packshot, the bottle, the interface, the face in the campaign is the actual anchor, and the model animates outward from it.

With two images, the pair is read as a first and last frame, which is useful for a placement that needs a defined opening and a defined end — a packshot that lands and settles, for instance. What this path cannot do is take a larger reference set: up to seven images, non-empty and required, is the base model's multi-reference behaviour, and it is the one to reach for when the shot must reconcile several angles of the same product.

Cut-downs are arithmetic, not taste

A campaign is usually several cut-downs of one idea: a 6-second bumper, a 15-second pre-roll, a 30-second cut, each in more than one aspect ratio. Since the reference bills generation on a per-second basis and prices Pro at roughly two times Turbo, the budget consequences are arithmetic and predictable.

The same beat rendered at six seconds costs a fraction of the same beat at fifteen; rendering the fifteen-second master in Pro and deriving the shorter cut-downs from it is the shape of the saving, provided your workflow can trim outside the generator. The alternative is rendering every variant at every length on the premium tier, which is the most common way an ad workflow quietly becomes the most expensive line in the account.

Iterate on Turbo, where the parameter surface is identical, and only pay the higher price on the frame that is actually placed.

Building a placement set across formats

The aspect ratio parameter takes the three common formats and, per the reference, the Q3 family supports arbitrary ratios beyond them, so unusual slots are not automatically a problem. One trap deserves naming: passing reference images often lets the image decide the ratio. If you supply a landscape packshot and leave the parameter unset, you may inherit a 16:9 composition while the buy is a vertical slot.

Set the ratio explicitly on every request in a campaign, and the same rule applies to vertical-first production — if the primary placement is 9:16, render the master vertical and cut wider versions from it, because re-rendering at the wrong ratio changes the composition rather than cropping it. Resolution behaves the same way across formats: 540p, 720p and 1080p, with 720p as the default.

The 5,000-character prompt is the script

Prompt is a required field with a 5000-character ceiling, and the reference asks that it describe subject, action, scene, camera and style. For a placement that budget is doing double duty, because audio defaults to true and the generated speech and background effects are derived from these same words. If the buy includes a voiceover or a jingle, the line and its direction belong in the prompt; there is no separate audio script field on this endpoint.

The reference also notes that subjects can be referenced with an `@name` placeholder when a subject has been established, which is how you keep referring to the same product across several lines without re-describing it. Practically, treat the prompt as a shot list with copy: shot, framing, motion, and what is heard, in that order.

When an ad brief is the wrong reason to pick this

Three cases where this is the wrong call, stated so they are not discovered late. First, if the placement needs fidelity above what the top tier produces, no variant in this family will get you there — the reference describes Pro as the highest tier available. Second, if the shot needs several reference images reconciled into one consistent identity, that is the base `viduq3` model's documented multi-reference behaviour, not a Pro or Turbo job; Pro and turbo accept a first frame or a first-and-last pair and stop there.

Third, if the work is a high volume of short variants with no single hero, Turbo is the correct primary and Pro is at most a finishing pass. And to be explicit about what this page is not: it is not evidence that Vidu ships an ad-tuned Q3 model, because the reference does not list one.

FAQ

Does Vidu publish a separate ad-specific Q3 model?
Not in the generation documentation, which lists exactly three Q3 identifiers: `viduq3-pro`, `viduq3-turbo` and `viduq3`. This page describes the commercial-placement brief and the family spec it is served from, and does not claim a fourth, ad-tuned tier.
How do I keep the product in frame for a placement?
Supply one reference image and it becomes the first frame, so the model animates from your actual asset instead of inventing an approximation. Two images are read as a first and last frame pair. Larger reference sets, up to seven and non-empty, belong to the base `viduq3` model.
Which model should I use for many short cut-downs?
Draft on Turbo, which the reference describes as fastest with the best cost-performance, and the same 1-to-16-second range as Pro. Generation is billed per second and Pro costs roughly two times Turbo, so paying the higher price on the single placed frame rather than on every variant is the main saving available.
How do I get a vertical 9:16 placement?
Set the aspect ratio explicitly to 9:16 on the request. The family supports arbitrary ratios beyond 16:9, 9:16 and 1:1, but when you pass reference images the ratio is often decided by the image, so leaving the parameter unset can quietly hand you a landscape composition for a vertical buy.

FAQ

Can I use Vidu Q3 Ad directly on this page?
Yes. The creator above is the working Ofrai studio with Vidu Q3 Ad 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 Ad?
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 Ad.

Commercial use depends on the selected model and model terms.