Google Launches Genie 4 for Interactive World Creation in Project Genie

Image: Google · Source
Google debuts Genie 4, an advanced interactive world model, now available for AI Ultra subscribers in Project Genie. It features real-time environment editing, native adaptive audio, and a community gallery.
Google has officially launched Genie 4, the latest iteration of its interactive world model, as part of the Project Genie platform. This new version is being rolled out to eligible Google AI Ultra subscribers and aims to significantly enhance the responsiveness, social interactivity, and creative control users have over generated digital worlds.
Genie 4 introduces several groundbreaking features that enable users to modify their virtual environments in real time using simple natural language prompts. For example, one can change the weather, add characters, or alter scene elements seamlessly while exploring a generated world. Additionally, Genie 4 integrates native audio that adapts dynamically to the unfolding scenes — from ambient soundscapes to character dialogue — bringing an immersive new layer of realism.
Alongside the new model, Google has updated the existing Genie 3 model to support extended exploration sessions of up to three minutes, panorama grounding using 360-degree images, and linked sequences of worlds for continuous journeys. These improvements provide users with new ways to navigate and interconnect rich virtual environments.
A notable addition is the Project Genie agent, an AI assistant that facilitates world creation through conversational interaction. Instead of relying on a single prompt, users can describe their vision naturally, enabling the agent to orchestrate Google's multimodal models to generate initial visuals and customized background scores.
Furthermore, Google has introduced a community gallery where users can publish their creations for public exploration and remixing. The gallery features a variety of curated and community-built worlds, such as “Villa Giallo,” a gothic manor mystery, and “The Life of Ants,” offering a photorealistic perspective of an ant colony. Users can also modify existing worlds directly in the gallery with the help of the Project Genie agent.
Google emphasizes its commitment to responsible deployment. All generated video and audio include SynthID watermarking, and safety screenings are applied to model inputs and outputs. Ongoing research addresses challenges like camera latency, complex physical interactions, and long-term scene consistency.
Project Genie, combining Google’s most advanced world and Gemini models, is accessible via a web browser, inviting AI Ultra subscribers to start creating, exploring, and sharing immersive interactive worlds today. Google AI Pro members can join a waitlist for future access.
This initiative reflects Google's continued investment in pushing the boundaries of AI-generated immersive content, enabling users to co-create complex digital universes with unprecedented ease and realism.
Sources and original reporting
Read the original source ↗

Comments (0)
No comments yet. Start the discussion.
Write a comment
Comments are published after moderation. Your name and comment will be visible publicly. Account