Google Deepmind unveils Genie 4, real-time world model with sound and dialogue

Google Deepmind's Genie 4 is a real-time world model that simulates interactive environments frame-by-frame with direct audio generation and character dialogue. Users navigate scenes via keyboard and text prompts, while Gemini models plan ahead to maintain coherence. The system handles persistent background audio, real-time character reactions, and user-controlled agents, though longer sessions currently struggle with consistency and object persistence.