Google DeepMind launched its newest AI system, the Genie 4 world model, in a research preview on Tuesday. TechCrunch reports that the software generates photorealistic interactive environments that users can alter mid-simulation by typing live prompts, while also producing synchronized ambient noises, music, and interactive dialogue.
Key takeaways
- Google DeepMind released Genie 4 in research preview on Tuesday.
- The system maintains consistent scenes for up to two minutes.
- Gemini integrates closely with Genie 4 to plan steps and check frames.
- Trusted testers are already exploring the software for film prototyping.
- Google plans to bring the model to AI Ultra subscribers and API partners soon.
What is the Genie 4 world model?
Built as the successor to Genie 3, the latest Google DeepMind release doubles the scene consistency limit of its predecessor up to two minutes. According to TechCrunch, it offers improved physics along with more humanlike characters. Users can execute free-form edits in real time, such as changing the camera angle or weather, or introducing new people and objects into the simulation.
The system works closely with Gemini, Google's AI model, which acts as a director to plan steps and check frames while the visuals render. In a research test tied to Gemini Robotics, a Gemini 4 Argon model split tasks into separate steps that Genie 4 simulated instantly. Tejas Kulkarni, a research director at Google, stated during a briefing that hybrid systems will combine Gemini writing code and text with Genie simulating reality.
How is AI video generation changing the industry?
The release arrives shortly after Runway launched its own playable second-generation world model, GWM Worlds 2, featuring dynamic responses and audio generation. Meanwhile, the industry is seeing consolidation as World Labs gets acquired by AMD, Amazon holds talks to buy Decart, and Rivian acquires Third Dimension for in-house simulation testing.
Potential use cases for Genie 4 include gaming, interactive learning, and robotics. Google also maintains Veo, an AI video-generation model that creates high-definition clips with synchronized audio, which filmmakers use frequently. Diego Rivas, a Google DeepMind engineer, noted that testers are pointing to a new use for Genie 4 in film prototyping and rapid ideation.
What happens next?
Genie 4 is currently available as a research preview for trusted testers. Google plans to roll out the model in the coming weeks to Google AI Ultra subscribers inside Project Genie, along with early access partners of Google AI Studio and the Gemini API. Rivas acknowledged that the software is still in early days, noting that objects can drift during longer runs and the team has not yet perfected audio accuracy or in-world text legibility.
Frequently asked questions
Google DeepMind released Genie 4 in a research preview on Tuesday, October 27, 2026, according to TechCrunch.
Genie 4 adds photorealistic graphics, real-time audio generation like music and dialogue, free-form edits, and support for maintaining consistent scenes for up to two minutes.
Gemini acts as a director by planning steps and checking frames while Genie creates the visuals, allowing hybrid workflows where Gemini handles text and code while Genie simulates reality.
Genie 4 is currently available as a research preview for trusted testers, with plans to roll out to Google AI Ultra subscribers and Gemini API partners in the coming weeks.
Pitched commercial use cases include gaming, interactive learning, robotics testing, and rapid ideation or film prototyping for creators.
Based on reporting by TechCrunch. Read the original article
