The announcement
Google announced Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking. Both are built for spoken conversations that can continue while the assistant handles a task. Google describes Live as the faster, lower-cost option and Extended Thinking as the one for more complex, multi-step work. The announcement was published September 15 and updated September 17.
What it means for you
If you want to talk through a task instead of typing it, this is the kind of model to watch. Google says access is rolling out through its API and AI Studio, with different availability across Search Live, Gemini Live, Workspace, and enterprise products. A model announcement does not mean every feature is available to every account today. Test it on a real conversation and check the product and plan you use before treating a benchmark claim as a personal speed or quality guarantee.