Google DeepMind Unveils Genie 4

On October 27, 2026, Google DeepMind introduced Genie 4, a research "world model" for real-time interactive generation of virtual worlds featuring video, physics, sound, music, and dialogue, currently available in a closed preview.

Compiled by Sergey KostenchukPublished 2026-10-27Updated 2026-10-27

2026-10-27 Research Google

🤖 Google DeepMind Unveils Genie 4

The research "world model" generates virtual worlds in real time: photorealistic video with physics, background sounds, music, and character dialogue. Text commands can change weather, camera angles, and add objects on the fly. Scene planning is handled by Gemini models.

🌍 For robotics and agentic systems, these worlds provide safe simulations for learning, while for game development and film, they offer a fast "draft mode": prototyping levels and previsualizing scenes.

👤 Access is currently restricted: a preview for trusted testers, followed by Google AI Ultra subscribers (Project Genie) and Gemini API and AI Studio partners. Visual drift, disappearing objects, or object changes may occur in long sessions.

Source 1: https://deepmind.google/blog/genie-4-a-new-generation-of-world-simulation/

Source 2: https://blog.google/innovation-and-ai/models-and-research/google-deepmind/introducing-genie-4/

Sources