- Home
- AI Video Generator
- Vidu Q4
Vidu Q4: Cinematic AI Video With 15 References, Native Audio and 4K
Vidu Q4 Preview is ShengShu's newest video model, released on October 7, 2026. It turns one image or up to 15 reference images into 3- to 16-second clips with sound, keeps voices consistent with reference audio, and renders up to 4K. Learn what it does, then create image-to-video and reference-to-video clips on Nereo with Seedance 2.5, Kling 3.0 and MiniMax H3.
Image to Video
Upload an image to animate
Vidu Q4 Features
Build One Scene From Many References
Reference-to-video takes 1 to 15 images. Tag them in the prompt, for example @1 and @2, and the model keeps each subject recognizable while placing them in a new shot together.
Reference-to-video: combine a character, a product, a pet and a location from separate images into one consistent cinematic frame.

Characters That Keep Their Voice
Add up to three MP3 reference clips of 3 to 12 seconds each. ShengShu positions this for more natural emotion and a consistent voice across a series, which matters for short dramas and recurring characters.
Reference audio: a dialogue scene where each character speaks in the voice from their reference sample, with sound generated together with the picture.

Fast Action Without Losing the Subject
The release focuses on high-dynamic scenes such as martial arts, running and confrontations, with motion that flows from start to finish and camera switches inside a single clip.
High-dynamic motion: fast, continuous action shots such as chases and leaps are one of the main improvements ShengShu describes.

Spectacle and Finishing Detail
Fireworks, particles and energy effects are rendered with finer detail, and 2K and 4K output give more room for grading and large screens.
Effects and resolution: VFX-heavy shots and up to 4K output for trailers, ads and cinematic intros.

How to Use Vidu Q4
Vidu Q4 Preview starts from images, not from a text prompt alone.
Choose Image or Reference Mode
Use one image as the first frame for image-to-video, or upload 1 to 15 reference images of your characters, products and places for reference-to-video. Text-only generation is not part of the preview.
Write the Shot and Add Voices
Describe the action, camera and dialogue. In reference mode, refer to subjects as @1, @2 and so on, and optionally add up to three 3- to 12-second MP3 voice samples.
Pick Length and Resolution
Set 3 to 16 seconds and 540p, 720p, 1080p, 2K or 4K. Draft at a lower resolution first; cost rises with resolution and length.
What Creators Use Vidu Q4 For
ShengShu aims Q4 at teams that produce video at volume.
Short Dramas and Series
Recurring characters with consistent faces and voices across episodes, using reference images and reference audio.
Ads With Real Products
Put the exact product, packaging and spokesperson from reference photos into new lifestyle scenes.
Action and Trailer Shots
High-motion chases, fights and effect-heavy moments for trailers, game promos and channel intros.
Music Videos
Performances and stylized scenes with native sound, then cut to the track in an editor.
Film Previsualization
Test blocking, camera moves and look from concept art before a shoot or a full render.
Social Content at Scale
Turn one set of references into many short clips for Reels, Shorts and TikTok.
Vidu Q4 Preview vs Vidu Q3 Pro
| Feature | Vidu Q4 Preview | Vidu Q3 Pro |
|---|---|---|
| Release | October 7, 2026 | Earlier in 2026 |
| Text to video | ||
| Image to video | ||
| Reference to video | 1–15 images + up to 3 voice clips | |
| Clip length | 3–16 seconds | 1–16 seconds |
| Max resolution | 4K | 1080p |
| Native audio | ||
| API price at 1080p | 24 credits per second | 24 credits per second |
Planning a Vidu Q4-Style Shot on Nereo
Vidu Q4 is not on Nereo yet; these tips apply to the reference-to-video models available today.
Start From Strong References
Clear, well-lit images of each character, product and location matter more than a long prompt.
Name Each Subject
Say which reference is which in the prompt so the model keeps identities straight.
Draft Low, Finish High
Test framing at a lower resolution and shorter length, then render the final take.
Plan Multi-Shot Stories
Lay out shots first with the AI Storyboard Generator, then generate each one with the same references.
Check the Credit Estimate
Each Nereo tool shows the cost for the selected model, length and resolution before you generate.
Mind Rights and Likeness
Only use images and voices you own or have permission to use.
Vidu Q4: Strengths and Practical Considerations
Pros
- Deep reference control — Up to 15 images plus up to three voice samples in one generation.
- Sound and picture together — Native audio removes a separate sound pass for many clips.
- High resolution ceiling — 2K and 4K output with 10-bit color reported at the top tiers.
Cons
- Preview limits — No text-only generation yet, and features or prices may change before the full release.
- 16-second clips — Longer pieces need several shots joined in an editor.
Vidu Q4 FAQ
Find answers to common questions about our AI video generation service.
Vidu Q4 is the newest video generation model from ShengShu Technology. The current release, Vidu Q4 Preview, came out on October 7, 2026 and supports image-to-video and reference-to-video with native audio, up to 15 reference images and up to 4K output.
The available version is called Vidu Q4 Preview. ShengShu released it ahead of the full Q4 model, so features, limits and prices may change when the final version ships.
Not in the preview. Vidu Q4 Preview supports image-to-video (one start frame) and reference-to-video (1 to 15 reference images). For prompt-only video, Vidu Q3 or another text-to-video model is needed.
Each clip can be 3 to 16 seconds long. The API default is 5 seconds at 720p.
540p, 720p, 1080p, 2K and 4K. The 2K and 4K tiers are reported to use 10-bit color.
On Vidu's own API, ShengShu lists Q4 Preview from about $0.045 per second at 540p up to about $0.39 per second at 4K, with a 30% launch discount on both modes through November 30, 2026. Prices may change after the preview.
At launch, Vidu Q4 Preview was reported to debut third on Artificial Analysis's image-to-video leaderboard. Leaderboards change often, so test it on your own shots before deciding.
Vidu Q4 is not in Nereo's model selector yet. The buttons on this page open Nereo's Image to Video and Reference to Video tools, where models such as Seedance 2.5 (up to 30 seconds and up to 30 reference images), Kling 3.0 and MiniMax H3 handle similar work today. Check the model list and credit estimate there before generating.
