Vidu AI

Overview

Vidu is the AI video generator from Shengshu Technology, a company spun out of Tsinghua University’s AI research labs. Its Q4 model, launched as a public preview on October 7, 2026, is built around lifelike performance: facial expression, emotion, body movement and voice are generated together so characters read as acting rather than being animated by a prompt. Vidu positions Q4 as its all-in-one video flagship, alongside Vidu S for real-time generation.

Key Features

• Reference to video: Combine up to 15 image references and up to three voice clips to keep characters, products and places consistent across a scene
• Image to video: Animate a single start frame from a text prompt, keeping the original image’s aspect ratio
• Native audio: Generate dialogue, sound effects and ambience with lip-sync in the same pass, without a separate audio step
• Up to 4K: Output at 540p, 720p, 1080p, 2K or 4K, with 10-bit color at the higher resolutions
• Camera control: Stable moves and cuts that follow the prompt, with automatic multi-shot switching inside one clip
• 3 to 16 second clips: Short takes with multi-shot structure, from the web app or the API

Use Cases

• Short drama and serialized storytelling with recurring characters
• Advertising and product video built around a reference image
• Social content that needs motion and sound in one generation
• Pre-visualization and concept clips for film and animation teams

Pricing

Official rates (Shengshu, October 2026):

• Free: every account gets trial credits, plus daily bonus credits for logging in
• API pay as you go: from $0.045 per second at 540p, scaling with resolution up to $0.39 per second at 4K
• Launch promotional pricing: from $0.014 per second, with image-to-video and reference-to-video generation at a 30% discount through November 30
• Web app: subscription plans for individuals and teams, on top of the pay-as-you-go API

Q4 Preview is released ahead of the finished Q4 model, and Shengshu says creator feedback on performance, creative control and production needs will inform the final release.