Vidu Q3
Native audio-video generation — 16 seconds of 1080p with dialogue, SFX, and lip-sync in a single pass.
About
Vidu Q3 is Shengshu Technology's flagship video model, unveiled January 30, 2026. It generates picture and sound together in one pass — dialogue with lip-sync, ambient effects, and background music matched to the scene — rather than adding audio in post-processing. Clips run up to 16 seconds at native 1080p and 24fps, with Smart Cuts multi-shot sequencing (automatic camera-angle changes inside one generation) and cinematic camera control. A Reference-to-Video mode added in April 2026 combines multiple reference inputs (subjects, environments, costumes, styles) for story-driven work.
The independent signal is strong: Artificial Analysis ranked it #2 globally on its video benchmark at launch (#1 in China), and it later took #1 on SuperCLUE's first global reference-to-video leaderboard. Shengshu reports 40M+ creators across 200+ countries. The model is built on the company's U-ViT architecture with its TurboDiffusion acceleration research.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
Independent: Artificial Analysis video benchmark #2 global at launch (#1 China); SuperCLUE reference-to-video leaderboard #1 (Apr 2026).
Provider: native 1080p, 16s duration claims.
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 7/10
Official site verified live (HTTP 200); independent Artificial Analysis and SuperCLUE rankings corroborate the quality claims — rare for video models. Docked for: (1) proprietary with no weights; (2) marketing copy leans on creator-count figures that can't be audited.