Google DeepMind releases Genie 4 world model with native audio
The gist
Google DeepMind released Genie 4, a world model that generates audio and interactive scenes, to trusted testers.
The preview adds real-time text edits and doubles scene consistency to two minutes.
How it unfolded
- Aug 5, 2025Google DeepMind announced Genie 3, described as its first world model allowing real-time interaction, generating dynamic worlds at 24 frames per second and staying consistent for a few minutes at 720p, as a limited research preview.
- Oct 27, 2026Google DeepMind released Genie 4 as a research preview to trusted testers, announced by Raia Hadsell, Diego Rivas and Hang Qi. It is natively multimodal, generating music, ambient noise and interactive dialogue alongside the simulated world.
- Oct 27, 2026Genie 4 allows freeform real-time edits via text prompts, such as changing the weather, moving the camera or adding people and objects, and holds a consistent scene for up to two minutes, double its predecessor's limit.