Google has introduced Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking. Its announcement, published September 15 and updated September 17, 2026, describes two models for voice conversations and tasks.
What changes
Google positions the standard model for scale and cost efficiency, while Extended Thinking targets more complex reasoning. The company says the models can continue conversations while tools perform tasks in the background.
Reported capabilities include visual context and automatic switching between 97 languages. These are Google’s descriptions of the models, rather than independent findings from Guava.
Where access is rolling out
Both models are rolling out through the Gemini API and Google AI Studio. Enterprise access is in private preview, with additional business availability planned.
Consumer destinations differ: Google lists Search Live for the standard model and Gemini Live for Extended Thinking. Workspace availability depends on the application and subscription.
The rollout therefore does not mean every feature is immediately available to every account.
