Shengshu releases Vidu Q4 Preview video model

Shengshu Technology has launched Vidu Q4 Preview. It is the first public preview of the company's next-generation flagship model for generating expressive video with audio. The model is available now through Vidu's web product and API platform.

The preview accepts up to 15 image references for defining characters, wardrobe, props, products and environments. It also takes up to 3 audio references to keep voices consistent and guide emotional delivery. Output resolutions run from 540p and 720p through 1080p, 2K and 4K, with 10-bit color depth at 2K and 4K. Shengshu says the model improves character performance, with better coordination of facial expression, body movement and voice. It also claims more coherent camera cuts in action sequences and better integration of effects such as explosions and particles.

Promotional launch pricing starts at $0.014 per second. The company claims users get up to five times as much output for the same budget "under comparable output specifications and billing conditions." It did not say what it was comparing against. Final pricing, resolutions and features may vary by plan and region.

"Advanced video models should prove themselves beyond carefully selected demos," said Yihang Luo, co-founder and CEO of Shengshu.

The announcement includes no benchmark results or independent evaluations. Shengshu says it is releasing the preview ahead of the full Q4 model so creator feedback can shape the final release. It has not given a date for the full release.