Google DeepMind unveiled its newest world model, Genie 4, as a research preview on October 27, 2026. This advanced generation of the AI world model creates highly photorealistic and interactive environments that include real-time audio such as music, ambient sounds, and conversations aligned with the scene. Genie 4 allows users to adjust features on the fly, like weather or camera angles, and can maintain scene consistency for up to two minutes, which is twice the duration of its predecessor, Genie 3. It also improves physical accuracy and renders more lifelike characters, offering a richer simulation experience.

Genie 4 tightly integrates with Google’s Gemini AI, which acts as a director by orchestrating planned steps and real-time frame checking during simulation. This hybrid approach enables sophisticated experiments, such as testing robot task sequences in a virtual environment before deployment in the real world. Google envisions Genie 4 supporting use cases in gaming, interactive education, and robotics, with early users also highlighting the model’s potential in rapid film prototyping and creative storyboarding due to its seamless real-time responsiveness.

The launch of Genie 4 occurs amid a rapidly consolidating world model sector. Following recent acquisitions and investments—including AMD’s purchase of Fei-Fei Li’s World Labs for $8.2 billion, Amazon’s planned acquisition of Decart, and Rivian’s takeover of Third Dimension—Google’s release represents a key move to maintain leadership in this competitive market. Genie 4 will initially be accessible to trusted testers and Google AI Ultra subscribers through Project Genie and early Gemini API and AI Studio partners, signaling a phased rollout targeting advanced users and developers.

Despite its advances, Google acknowledges that Genie 4 is still evolving, with some challenges remaining in keeping scene details stable over longer durations and ensuring audio and in-world text quality. The company is actively gathering feedback from initial users as it works to refine these aspects. As the model becomes more widely available, Google anticipates that continued improvements and integration with other AI tools like Veo and Gemini Omni will further enhance its capabilities and creative potential for various industries.