Google has announced the release of Genie 4, a new world model that brings significant advancements in real-time interactive simulation. According to the company, Genie 4 offers improved physics, more humanlike characters, and the ability to generate native audio on the fly, creating a more immersive experience.
The model also allows for free-form edits, such as changing the weather or camera angle in real time, using natural language inputs. Genie 4 incorporates capabilities from the Gemini family, which provides complex reasoning and helps maintain simulation coherence and continuity.
The rollout of Genie 4 has started with a limited group of trusted testers, with plans to expand to Google AI Ultra subscribers and early access partners in the future. The introduction of Genie 4 marks a significant step towards achieving artificial general intelligence, enabling safer training environments for robots and agents, and enhancing creative tools across various industries.
Evidence lens
2 cited sources are linked to this briefing.