Google DeepMind Unveils Genie 4: Real-Time Interactive World Model Built on Gemini
Google DeepMind · youtube · 2026-10-28
Google DeepMind has announced Genie 4, described as its most realistic and interactive world model yet. Built on top of Gemini, Genie 4 lets users step inside dynamic 3D worlds in real time, navigate with standard keyboard controls, and steer scenes on the fly with natural-language prompts — the people in the world then move, react, and speak accordingly, with sound.
Genie 4 is rolling out as a research preview to trusted testers, and is coming soon to Google AI Ultra subscribers via Project Genie, plus early access partners of the Gemini API and Google AI Studio. All footage in the announcement video is real-time interaction recorded from the model.
Related event: DeepMind Unveils Genie 4, Its Most Realistic Interactive World Model(31 posts)→
More from Models
- Liquid AI launches LFM3 family: 6 compact models plus d1 decision models in a single pass — antoine_chaffin · 2026-10-28
- Follow-up: GPT-5.6 terra would be a hit if OpenAI cut its price 80% instead of 20% — JoshPurtell · 2026-10-28
- Dev calls out OpenAI's messy GPT-5.6 pricing: mini/nano hiked, luna reverted, terra left behind — JoshPurtell · 2026-10-28
- Liquid AI LFM3's hybrid architecture: 6 global attention layers, 256K context, 25% faster inference — cneuralnetwork · 2026-10-28
- Dev names GLM 5.5 Flash and Qwen 4 27B as the models he's most excited about — victormustar · 2026-10-28
- Maxime Labonne demos d1: bounding boxes for named objects in a single forward pass — maximelabonne · 2026-10-28