Vidu Q2 dropped in September 2025. It's from Shengshu Tech tied to Tsinghua University.

This new model turns images or references into short videos. It leans into tiny facial moves, smooth zooms, and quicker results. You can make clips from 2 to 8 seconds, lock in the start or end frame.

Compared to Q1, Q2 brings better facial moves, steadier shots, and sticks to prompts better, with less rubbery movement.

Pro (Cinematic) mode focuses on longer, movie-like scenes with rich visuals. Turbo (Flash) mode makes fast, motion-heavy clips.

Vidu Q2 Cinematic Examples

Done many variations of this prompt but this is the closest I could get to a realistic punch. Mostly his head doesn't move at all on impact and the fist barely touches him. Generated on October 16, 2025

for link to original.

Compare Models

Vidu can do multi-shots! Here it does zoom first, then cut. In another try it cut both ways but I like this output better overall. Generated on October 13, 2025

for link to original.

Compare Models