Open the workspace on the tier it starts with: Vidu Q3 Turbo runs text, image and first/last frame tasks, and carries the lowest per-second official rate of the three Q3 tiers at 1080p.
Vidu Q3 Turbo is the model this workspace loads by default: text, image or first/last frame input, 2 to 15 whole seconds, and 720p or 1080p output.
This model does not support reference video input
Generated content will appear here
Facts reviewed:
Vidu documents more routes for this model than this page opens. Both lists are shown so the gap is visible rather than implied.
This is the widest gap between official documentation and local controls anywhere in the Vidu Q3 line, and it is worth stating plainly: Vidu documents four generation routes for viduq3-turbo, and this page opens three of them. Reference to video has its own published per-resolution rates and its own API page, so it is not a minor omission — it is a route the model genuinely supports and this platform does not expose. The same applies to the audio parameter, which Vidu restricts to Q3 models and which has no control here, and to off-peak scheduling, which this route does not offer at half rate. What this page does run is the three routes listed above at 2 to 15 seconds and 720p or 1080p, and the point estimate shown in the workspace is the only current statement of what a task costs here.
Vidu Q3 is the third generation of Vidu video models, and the official model map groups five models under that label. Vidu Q3 Turbo carries the lowest per-second credit cost of the three Q3 tiers priced for general use — 13 credits per second at 1080p against 24 for Pro and 25 for Pro Fast, read from the same official pricing table. It is also the tier this platform loads by default when you open the video workspace, which makes it the route most tasks will actually run on.
The other thing the official documents say about viduq3-turbo is that it supports more routes than its siblings. The model map marks image-to-video, reference-to-video, first/last frame and text-to-video as available for this model, while viduq3-mix, viduq3-drama and viduq3-ad are reference-only and viduq3-pro carries three of the four. So the official picture is a tier that trades some per-second rate for the broadest route coverage in the series, described in the table with a single line: fast generation speed.
That breadth is where official documentation and this page diverge, and the divergence is the subject of the comparison section below. This page runs three of the four documented routes. Reference to video is documented for this exact model, priced in its own rows on the official table, and not selectable here. Audio generation, which Vidu lists as a Q3-only parameter, has no control here either, and neither does off-peak scheduling. Reading the official pages alone would leave you expecting all of those, which is precisely why they are listed as unavailable rather than quietly omitted.
Everything below is a setting or route this page actually exposes. Documented-but-absent capabilities are listed separately in the comparison above.
Write the scene, action, camera behaviour, lighting and pacing into one prompt and generate from text alone. Vidu prices text2video for this model in the same row as its other two main routes, so this is a priced route rather than an inferred one.
Upload one image you are authorized to use and describe how the subject, camera and environment develop away from that starting frame. It is the most commonly available route across the whole Q3 line.
Two visually compatible images define both the opening state and the destination of one continuous move. Useful whenever the shot has to land on a specific final composition rather than wherever the model ends up.
At 1080p the official table lists 13 credits per second for Turbo against 24 for Pro and 25 for Pro Fast. That single difference is what makes iterative shooting practical on this tier rather than an occasional indulgence.
Choose any whole number of seconds from 2 to 15. The official range is 1 to 16, so both ends are narrower here, and the local range is stated separately rather than presented as the official one.
Pick the resolution that matches the delivery surface. The official table prices 720p at 11 credits per second against 13 at 1080p, so the lower setting is a real cost lever rather than only a quality setting.
16:9, 9:16, 3:4, 4:3 and 1:1. Vidu documents 3:4 and 4:3 as supported specifically on Q2 and Q3 models, so the two narrower ratios are legitimately available on this tier.
The workspace recalculates the estimate for the exact combination you selected, updating as you change duration or resolution. It is the only current price signal for this route.
Everything below is a setting or route this page actually exposes. Documented-but-absent capabilities are listed separately in the comparison above.
The lowest official per-second rate makes repeated short attempts affordable, which changes the working method: instead of one carefully specified long take, run several variations and keep the one that lands.
Composition, framing and motion can be settled on this tier and then re-rendered elsewhere if the final delivery needs a different route. The draft costs a fraction of the final attempt.
Short vertical clips at 9:16 are usually watched small, so 720p is often sufficient and costs less per second on the official table. Choose the resolution from the viewing size rather than from habit.
Because cost scales with duration, a cheaper per-second rate buys more seconds for the same spend. Fifteen second clips are reachable here, but the official 16 second ceiling is not.
The first/last frame route is the least prompt-sensitive of the three available, which makes it the right choice when a client has already approved the opening and closing images.
Run the same scene with two different camera descriptions and compare the results side by side. At this rate the comparison costs little, and it answers questions that written documentation cannot.
Written for a tier that runs three routes and no audio, off-peak or reference controls, so nothing below depends on a setting this page does not expose.
A short text-to-video starter built to test how much motion the tier adds on its own when the prompt asks for very little.
A quiet street at dusk after rain. One person walks away from the camera at a steady pace. Reflections hold on the wet asphalt. Camera follows at walking speed, no cuts. Warm shop light spills from the left. Keep the figure consistent and keep the frame free of text and signage.
A framing test where the only camera instruction is a single continuous move.
Start from the provided image and push the camera slowly toward the subject over the whole clip. Lighting and background stay unchanged. Movement should be even from first frame to last, with no sudden change in framing or focus.
A first/last frame task that tests whether the tier can hold geometry across a single continuous move.
Begin on the first frame and finish on the second. Move the camera through the doorway into the lit interior at an even pace. Keep the wall angles straight and the light transition gradual, with no cut in the middle.
An image-to-video starter that makes any unwanted change to the product easy to spot.
Using the supplied image, rotate the camera around the product by about 20 degrees. The label stays readable and faces the camera throughout. Surface and backdrop remain identical. No change to colour, proportion or material.
A text-to-video starter designed to be run twice with a single line changed, so the two results can be compared directly.
A cyclist rides along a coastal road at golden hour. Camera option A: a slow lateral tracking shot moving with the rider. Keep everything else in the description identical so the two renders differ only in camera behaviour.
A scene where the intended change is environmental, which isolates camera and object stability.
A fixed wide shot of an empty room with a large window. Daylight through the window gradually warms from midday white to late afternoon amber, and shadows stretch slowly across the floor. Camera does not move and no objects enter or leave the frame.
Left column reads what the official Vidu documents list for viduq3-turbo specifically; right column reads what this page actually exposes.
| Route or setting | Documented for viduq3-turbo | Available on this page |
|---|---|---|
| Text to video | Listed as text2video on the official pricing table | Enabled |
| Image to video | Listed as img2video on the official pricing table | Enabled |
| First and last frame | Listed on the official model map | Enabled |
| Reference to video | Listed as reference2video with its own published rates | Not opened here |
| Audio parameter | Supported on Q3 models only | No control on this route |
| Off-peak scheduling | Half rate, generated within 48 hours | No control on this route |
| Resolution range | 540p, 720p, 1080p | 720p, 1080p |
| Duration range | 1 to 16 seconds | 2 to 15 whole seconds |
| Aspect ratios | Five ratios including 3:4 and 4:3 | Five ratios including 3:4 and 4:3 |
| Official rate at 1080P | 13 credits per second, $0.065 | Point estimate shown before submission |
| Official rate at 720P | 11 credits per second, $0.055 | Point estimate shown before submission |
| Difference from Pro at 1080P | 47 percent lower than Pro | The tier the workspace loads by default |
Vidu prices its own API per second, with one credit marked at $0.005 on the official pricing page. The workspace estimate is this platform's separate schedule, so use the official figures to judge the tier and the on-screen estimate to judge the task.
View platform pricingThe estimate recalculates for the exact settings you selected and moves with both duration and resolution. Because this tier sits at the lowest official per-second rate of the three, the same spend buys more seconds here than on Pro — which is the practical reason to settle composition on this route before moving elsewhere.
The official table lists 11 credits per second at 720p and 13 at 1080p for this tier, before off-peak is applied. For vertical social delivery watched on a phone, 720p is frequently sufficient and costs less per second; for anything that will be viewed large or re-graded, the higher setting is the one to budget for.
The workspace loads with Vidu Q3 Turbo already selected. If the brief needs a different route or a different per-second rate, switch model before you write the prompt rather than after.
Text for a scene you are describing from scratch, one image when the subject or product must be preserved, or a first and last frame pair when the shot has to end on a specific composition.
Whole seconds from 2 to 15, then 720p or 1080p. Both settings move the estimate, and resolution is the cheaper variable to adjust when a draft is still being explored.
There is no movement control on this tier and Vidu documents the movement amplitude parameter as having no effect on Q3 models, so shot size, direction and speed belong in the prompt text.
Review the settings and the point estimate together before submitting. The estimate reflects the exact combination on screen, so it changes the moment any field does.
The official description of this model is a speed statement, not a quality claim. Vidu labels viduq3-turbo with fast generation speed and does not publish a comparative output-quality verdict against the other Q3 tiers, so nothing on this page asserts one. Where a brief depends on a specific look rather than on turnaround, test the same prompt on more than one tier before committing a full sequence to either.
Four documented capabilities are absent from this route and each one changes what a project can plan for. Reference to video, which Vidu documents for this exact model with its own published rates, is not selectable here. Audio generation, a Q3-only parameter in the official API, has no control. Off-peak scheduling at half rate is unavailable. And the resolution pool omits 540p while duration stops at 15 seconds rather than 16. If any of those four is a requirement rather than a preference, generate from a route that opens it.
Figures on this page describe how Vidu prices its own API and are not this platform's point prices. No promotional pricing is reproduced here, because the official Q3 section of the pricing table contains no discount entries. Rates and route availability can change without notice, so verify the official table before those numbers enter a budget or a client quote.
Generated video is not deterministic and can contain unstable identities between shots, unintended cuts, mismatched dialogue, inconsistent props and imperfect audio alignment where audio applies. Review the full clip at normal speed, then re-check the opening and closing two seconds and every change of shot size. Upload only material you are authorized to use, avoid deceptive or privacy-invasive output, and keep prompts, permissions and settings with the approved result.
Routes, capabilities and rates can change. Official Vidu documentation and platform information are listed separately below, and every official page listed here was reachable without a login when this page was reviewed.
Return to the workspace, set the duration and resolution, and confirm the point estimate before submitting a task on the tier that loads by default.