ARTIFICE DAILY AI + CYBER SECURITY BRIEFING

Artificial Intelligence

Google launches Genie 4 world model

Google said Genie 4 would generate interactive AI experiences with more realistic worlds, audio and longer simulations, and it would first reach trusted testers as a research preview.

The Verge

Google announced Genie 4 today as a world model for generating interactive experiences with AI. The company said it introduced four main changes: more realistic worlds, the ability to revise an experience in real time with a natural language prompt, audio generation and Gemini 3.8 Flash underneath the simulations.

Google said Genie 4 could produce experiences of about two minutes, rather than the one-minute limit of Genie 3. It also said the model could be useful for pre-visualisation in filmmaking, prototyping game ideas and making training scenarios for robots.

The company said longer simulations could lose detail and object permanence over time. It is launching Genie 4 first as a research preview for trusted testers, and Google said it had opened a waitlist for access through Project Genie.

Google said Genie 4 would be coming soon to Google AI Ultra subscribers. The article said the writer had not used the new model personally and could not compare it directly with the earlier Project Genie preview.

Named in this story

Companies

Google
announced the Genie 4 world model

Products and systems

Genie 4
the world model for generating interactive experiences
Gemini 3.8 Flash
provided the underlying model used by Genie 4
Genie 3
the previous model that could generate one-minute experiences
Project Genie
the preview through which trusted testers could try Genie 4
Google AI Ultra
the subscription tier that would get Genie 4 soon

How the source tells it

The piece was a routine product announcement with light promotional framing from Google and a brief caution about limits.

  • vendor boosterism Google’s own description and intended use cases carried the main persuasive weight
  • novelty the launch was presented as a new model with several added capabilities and a longer output window