Google DeepMind made its latest world model, Genie 4, available as a research preview on Tuesday. The system produces the most photorealistic and interactive environments to date, letting users type prompts during a simulation to alter the scene, while generating matching real-time audio such as music, ambient sounds and dialogue.
The launch follows AI video startup Runway's release of GWM Worlds 2, its second world model that also reacts dynamically to user input and creates audio. At the same time, the sector is seeing consolidation: AMD is acquiring Fei-Fei Li's World Labs, Amazon is negotiating to purchase Decart, and Rivian has bought Third Dimension to add simulation tools for self-driving tests.
Genie 4 represents Google's latest effort to stay ahead in a crowded market.
Google introduced the predecessor Genie 3 as a research preview a little over a year ago. Compared with that version, Genie 4 delivers far higher photorealism, includes audio, and supports freeform edits such as changing weather, camera angle, or adding people and objects in real time. It can keep a consistent scene for up to two minutes, twice the previous limit, and offers improved physics and more human-like characters.
Gemini, Google's AI model, works more closely with Genie 4, acting as a director that plans steps and checks frames while Genie generates visuals. In a research test linked to Gemini Robotics, a Gemini 4 Argon model broke tasks into steps that Genie 4 simulated live, showing a possible way to test robot scenarios in simulation before real-world execution.
"What we'll increasingly see in the future are these hybrid world models, where you have a combination of Gemini writing code and text, and Genie ingesting that in real time and simulating reality at different levels of abstraction," said Tejas Kulkarni, research director at Google, during a briefing.
Commercial use cases pitched for world models include gaming, interactive learning and robotics. One speculative interactive-learning example involves entering a painting of a historic scene and conversing with its subjects about the depicted event.
Trusted testers - the early-access users of Google's cutting-edge products - suggest another possibility: film prototyping.
"The real time aspect becomes most valuable in the areas where keeping a state of flow in the creativity space becomes incredibly useful," said Diego Rivas, a Google DeepMind engineer, at a press briefing. "For example, in the ideation phase, in the storyboarding phase, anytime you're kind of in need of rapid ideation."
Google also offers Veo, its AI video generation model that creates high-definition clips with synchronized audio from text, image or video prompts, a tool gaining traction among filmmakers.
Rivas added that users might eventually rely on Gemini Omni, a multimodal large language model capable of generating video from images, audio and text, to turn prototypes into creative assets, though he noted those abilities are still some way off.
Throughout the briefing, Rivas emphasized that Genie 4 is still in early stages and the final product experience remains unclear. The team acknowledged gaps: details and objects can drift in longer simulations, and audio accuracy and in-world text legibility are not yet fully resolved.
Genie 4 is now available as a research preview for trusted testers. In the coming weeks it will be rolled out to Google AI Ultra subscribers in Project Genie, Google's tool for generating interactive worlds from text or image prompts, as well as early-access partners of the Gemini API and Google AI Studio.
