Overview
Vidu Q4 Preview is the first preview build of Shengshu Technology's next-generation flagship video generation model. It was released on October 8, 2026 and rolled out to all users on day one. This generation concentrates its upgrades on three fronts: character performance, camera language and complex visual effects. Character emotion, motion and camera rhythm are matched more tightly, shots carry more tension, and complex visual effects come out with a higher degree of finish.
The most practical change sits on the input side. The model accepts up to 15 reference images and 3 reference audio tracks, which gives creators much stronger control over character appearance, scene atmosphere, costume, props and sound style. On the output side it covers 2K and 4K resolution, enough for professional delivery.
Price is the other signal worth noting. The launch price starts at 0.09 yuan per second, and the official claim is that under comparable specifications and billing terms, the same budget yields roughly five times as much video. Luo Yihang, co-founder and CEO of Shengshu Technology, frames this as converting model efficiency into a lower barrier to entry: video creation always involves retries, and a lower cost buys the freedom to shoot one more take or cut one more version.
Shengshu's core research team comes out of Tsinghua University. The team proposed the U-ViT architecture in 2022, and Vidu made its first public appearance in April 2024. The current lineup splits into two lines: the Q series for offline high-quality finished video, running from Q3 to Q4, and the S series for real-time interaction and real-time editing, running from S1 to S2.
Key Features
- Upgraded character performance: Character emotion, motion and camera rhythm are matched more tightly so performance carries the narrative
- More expressive camera language: Better shot expression and staging, giving footage a stronger sense of storytelling
- Complex visual effects: Complex visual effects come through with higher stability and finish
- Up to 15 reference images: Feed in up to 15 reference images in a single job to control character, scene, costume and props separately
- Up to 3 reference audio tracks: Three reference audio tracks can be supplied to pin down sound style and rhythm
- 2K and 4K output: Output covers 2K and 4K resolution, meeting the quality bar for professional delivery
- From 0.09 yuan per second: Launch pricing starts at 0.09 yuan per second, with the same budget yielding roughly five times as much video
- Fully available at launch: The preview build opened to everyone on release day, with no early access application
Use Cases
- Storyboard previews, character blocking and shot validation for short dramas and comic dramas, cutting trial and error costs before the real shoot
- Advertising and e-commerce teams generating and testing multiple creative variants across products, channels and audiences
- Serial content that needs a consistent character look and costume, held together by combinations of reference images
- Deliverables with a hard quality requirement at 2K or 4K
- Audiovisual work where sound style or rhythm needs to be specified
- Individual creators and small teams working with a limited budget who still want to iterate frequently
Pros
- The combination of 15 reference images and 3 reference audio tracks gives noticeably stronger control over character and scene
- 2K and 4K output covers professional delivery
- Pricing from 0.09 yuan per second brings flagship video generation down into cent-level territory
- The preview build opened to all users on day one, with no queue for access
- A Tsinghua-rooted team that started from the U-ViT architecture, with a clear technical track record
Pricing
The published launch price starts at 0.09 yuan per second, with several output tiers available; the unit price varies by resolution and model version. Under comparable specifications and billing terms, the same budget can generate roughly five times as much video. Exact pricing and current quotas follow the announcements on vidu.com and vidu.cn.
Summary
Vidu Q4 Preview is a flagship-tier update to the domestic video generation field, released on October 8, 2026. Character performance, camera language and complex visual effects all move forward together, joined by input freedom of up to 15 reference images and 3 reference audio tracks, output at 2K and 4K, and a launch price starting at 0.09 yuan per second. If you produce short dramas, comic dramas, or advertising and e-commerce assets and need to iterate often on a tight budget, this generation's mix of control and cost is worth testing first. If what you need is real-time interaction or editing a live video stream, the Vidu S2 line is the better match. If you are already on Vidu Q3, Q4 is the next flagship in the same series: more capable and cheaper, with almost no barrier to moving over.
Version History
- Vidu Q4 Preview released with upgrades to character performance, camera language and complex visual effects, plus 2K and 4K output (2026-10-08): Shengshu Technology released Vidu Q4 Preview, the first preview build of its next-generation flagship video generation model, on October 8, 2026, and made it available to all users at the same time. The generation upgrades character performance, camera language and complex visual effects, accepts up to 15 reference images and 3 reference audio tracks in a single job, and outputs at 2K and 4K resolution. Launch pricing starts at 0.09 yuan per second, with the company stating that the same budget can generate roughly five times as much video under comparable specifications and billing terms. Alongside the release the company opened the Q4 Preview Boundless Vision creation contest, with submissions accepted from October 8 to November 8, 2026.