Vidu Q2
Reference-to-video generation — blend up to 7 images into consistent, emotive clips.
About
Vidu Q2 is the flagship video model from Beijing-based ShengShu Technology, launched September 2025 with a 'Reference-to-Video' update that October. Its differentiator is Multiple-Entity Consistency: you feed up to seven reference images (faces, outfits, props, scenes) and it blends them into a single coherent video while keeping each element's identity intact. The model emphasizes expressive human performance — nuanced facial micro-expressions, natural body motion, smarter camera work — and has since expanded into a full GenAI stack with text-to-image and image editing. A global MaaS API lets businesses integrate it, and pricing undercuts Sora 2 and Veo 3.1.
The honest picture: as a Chinese vendor, ShengShu raises the usual data-governance questions for enterprise buyers. Independent benchmark coverage is thinner than for US competitors, so quality claims rest largely on the company's own demos. The free unlimited 1080p offer was a limited 2025 promotion — expect paid tiers for serious use.
What it can do
- Reference-to-video with up to 7 reference images
- Multiple-Entity Consistency across characters and props
- Expressive facial and emotional performance
- Text-to-image and image editing in the same model
- Cinematic camera movements and depth of field
- Global MaaS API with lip-sync and templates
Setup
Use at vidu.studio in the browser, or integrate via the Vidu MaaS API.
Prerequisites
An account; API use is pay-per-generation.
Use cases
Works with
Trust score — 8/10
Scored 8/10 (verified tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 57; Trending score 82/100 from researcher assessment.
Sources
Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.