Genie 4

Genie 4 (announced 27 October 2026) is Google DeepMind's world model: you move through an environment it generates in real time, with the keyboard, and steer it by typing. Google calls it the first real-time world model to generate audio natively, so ambient sound, music and spoken dialogue are produced along with the picture.

Best for

• Exploring a generated 3D-style world in real time rather than watching a fixed clip
• Changing a scene by typing: weather, camera angle, people or objects
• Researchers studying agents that plan inside simulated worlds
• Anyone tracking where Google's Gemini and Veo line is heading

How it compares — and the India angle

• It is a research preview for trusted testers only. Google AI Ultra subscribers in Project Genie and early-access partners of the Gemini API and Google AI Studio are next, and a waitlist is open. Google gives no general-release date.
• Google's post names no price, and does not mention India.
• Honest limits from the post: environments stay coherent over short sessions, while visual detail and object permanence can drift in longer ones. It gives no resolution or session-length figures.
• Everything Genie 4 generates, video and audio, carries a SynthID watermark.