# Google DeepMind Previews Genie 4, a Real-Time World Model That Generates Sound and Dialogue

Genie 4 keeps a scene consistent for up to two minutes, double Genie 3's limit, and is going first to trusted testers, Google AI Ultra subscribers and early Gemini API partners.

- Published: 2026-10-28T06:21:21.801Z
- Canonical: https://polylog.news/ai/2026-10-28/google-deepmind-previews-genie-4-a-real-time-world-model-tha
- Publisher: Polylog (AI desk)
- Section: tech
- Sources: [Google DeepMind (via Hacker News)](https://deepmind.google/blog/genie-4-a-new-generation-of-world-simulation/), [Polylog editors](https://polylog.news), [arXiv cs.LG](https://arxiv.org/abs/2610.35877)

Google DeepMind announced Genie 4 on October 27 as its new model for real-time interactive simulation. Unlike its predecessor, Genie 4 generates audio natively. Each world carries ambient sound that fits its setting, along with music, spati…

This story is for subscribers. Read it in full at https://polylog.news/ai/2026-10-28/google-deepmind-previews-genie-4-a-real-time-world-model-tha (subscription information: https://polylog.news/pricing).