⚡Google DeepMind announced Genie 4, a world simulation model that adds audio, more realistic physics, and real-time steering.

What it means

Google DeepMind has released Genie 4, its latest world simulation model that generates interactive, real-time 3D environments from text or image prompts. It builds on previous Genie models and integrates Google's Gemini multimodal AI to add native ambient audio, dialogue, and planning capabilities, making the simulated worlds more realistic and interactive than earlier versions.

Accuracy & sourcesSourced

This is an official announcement from Google DeepMind's blog, the primary source, so the claims are authoritative and include specifics about capabilities and limitations. Multiple secondary sources (Google Blog, YouTube, tech outlets) corroborate the release details.

How new

Genuinely new: the announcement and associated materials were published about 2 days ago (as of the search results date).

Go deeper

• Ars Technica — first to publish