Google is releasing Genie 4, an updated “world model,” to eligible Google AI Ultra subscribers through Project Genie. Outlets describe the update as allowing users to change environments within a single session using prompts, with added real-time “native audio” and dialogue-like audio capabilities.
Reporting also points to enhancements around how users create and share. Across sources, Genie 4 is presented alongside a new creation-focused Project Genie agent and a community gallery where users can build, remix, and share interactive worlds. Times of India additionally details that Genie 3 support includes longer sessions, world chaining, and 360-degree panorama features, helping explain the continuity between versions.
While the coverage largely agrees on the rollout path and core capabilities—mid-session environment changes plus native audio—the outlets vary in emphasis. Times of India concentrates on a feature-by-feature comparison between Genie 4 and Genie 3, while The Next Web and Decrypt focus more on the practical access for Ultra subscribers and real-time world building.
All sources focus on the product update and user-facing features; none describe broader policy, pricing beyond the subscription eligibility, or third-party testing results.