Google DeepMind releases Genie 4 world model in research preview
Original title: Google’s Genie 4 world model adds audio, more realistic worlds, and real-time steering
Trusted testers get it now; Google AI Ultra and Gemini API partners follow in the coming weeks.
Google DeepMind released Genie 4, a research-preview world model that adds matched audio, freer mid-simulation edits, and scenes it says stay consistent for up to two minutes. Gemini plans steps and checks frames while Genie renders visuals, including a Gemini Robotics experiment with Gemini 4 Argon. Google says details can drift, audio accuracy and in-world text remain weak, and the end product is still unclear.
Unconfirmed: Google says the finished product experience is not yet defined.
- Scene consistency is up to two minutes, double Genie 3’s limit.
- Gemini acts as director while Genie generates the visuals.
- Rollout later reaches Project Genie for Google AI Ultra subscribers.
- Engineers say film prototyping is an early use among testers.
- Audio accuracy and legible in-world text are still unsolved.
- Vendors
- Products
- Project GenieGeminiVeo
- Organizations
- Google DeepMind
- AI models
- Genie 4Gemini 4 ArgonGemini Omni
From the original article
3 min read at techcrunch.com
Google DeepMind released Genie 4, a world model with real-time audio, more realistic simulations, and mid-simulation text prompts, in research preview for trusted testers.
By Rebecca Bellan · techcrunch.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.