AI Research Atlas

Veo

Google DeepMind · 14 May 2024

Google's first Veo text-to-video model makes 1080p clips that can run beyond a minute, via VideoFX waitlist.

Understands cinematic terms (timelapse, aerial shots) and keeps motion coherent. Outputs carry SynthID watermarks. Announced at I/O 2024 for VideoFX waitlist users.

Date
Tuesday, 14 May 2024
Lab
Google DeepMind
Kind
model
Access
research preview

Waitlist access in VideoFX at announcement.

Sources

  1. blog.google/technology/ai/google-generative-ai-veo-imagen-3/

This record was checked against its sources on 6 October 2026. How we check