Google DeepMind says the model can track scene changes and help simulate robot tasks, but it is opening access through a research preview first.
By [Ryan Merket](https://runtimewire.com/author/ryan-merket)
· Published
Primary source: [x.com](https://x.com/Google/status/2115114426827501782)
Why it matters #
Genie 4's important test is whether Gemini reasoning can keep simulated scenes coherent enough for useful robot and agent training, beyond producing more convincing interactive demos.
Google DeepMind introduced Genie 4 on October 27th, adding live text controls and generated sound to its interactive world model as it tests whether simulated environments can help train or evaluate robots and AI agents. The company says the model can change a scene in response to prompts, maintain a record of what has happened and generate ambient audio, music and dialogue alongside the visuals. Its announcement on X describes the system as a research preview, with access initially limited to trusted testers.…
When you purchase through links in our articles, we may earn a small commission. This doesn’t affect our editorial independence.