World models
Source reports covering World models. Stories will appear when analysis and community reactions are ready.
2 reports · page 1 of 1
Connected topics are temporarily unavailable. You can still browse stories below.
Source reports
Google’s Genie 4 world model adds audio, more realistic worlds, and real-time steering
news
3 min read
Google DeepMind released Genie 4, a world model that generates interactive, photorealistic simulations with synchronized real-time audio, as a research preview on Tuesday. Users can alter a running simulation through text prompts, including changing weather or camera angles and adding people or objects. Genie 4 can generate music, ambient sound, and interactive dialogue, maintain a consistent scene for up to two minutes—twice Genie 3’s limit—and provides improved physics and more human-like characters. Google has integrated Gemini more closely with Genie 4: Gemini can plan tasks and check frames while Genie generates simulated visuals. In an experiment connected to Gemini Robotics, Gemini 4 Argon divided a task into steps that Genie 4 simulated in real time, indicating a possible use for testing robot scenarios before real-world deployment. Google cites gaming, interactive learning, robotics, and creative prototyping as potential applications. Genie 4 is currently available to trusted testers. Google plans to make it available in coming weeks to Google AI Ultra subscribers through Project Genie, as well as early-access Gemini API and Google AI Studio partners. Google DeepMind says the model still has limitations: visual details and objects can drift during longer simulations, audio accuracy remains unresolved, and in-world text may not be legible.
DeepMind unveils Genie 4 world model promising more realistic physics and humans
news
2 min read
Google DeepMind has introduced Genie 4, an updated world model that generates explorable virtual environments frame by frame in real time from user prompts. The system can create experiences resembling video games without pre-rendered environments or individually programmed interactive elements. DeepMind positions world models as a component of its work toward artificial general intelligence and says they could provide training environments for physical robots and support rapid creative prototyping. Genie 4 uses Gemini 3.8 Flash as a guiding agent that breaks down goals and evaluates generated frames, helping simulations retain context as users move through different environments. DeepMind says the model improves on Genie 3 in several areas, including more useful simulations, but acknowledges that generated worlds remain consistent for only a couple of minutes before losing track of their initial state. Users can alter a simulation while it is running by changing weather, adding elements, or moving the camera; Gemini helps steer the generated world after each update. Google DeepMind says Genie 4’s interactive generation does not require environments to be created in advance, though access may depend on paying Google.