Google launched two real-time voice models, Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking, for developers building voice agents and for its own apps. The standard Live model is the cheaper option for high-volume use, and it can respond to what a camera sees, switch among 97 languages mid-conversation, and run tools in the background while it keeps talking. Extended Thinking handles harder, multi-step work by reasoning as it speaks and narrating its progress on background tasks. Both are available now in the Gemini API and Google AI Studio, and Gemini Enterprise customers can try them in a private preview. Extended Thinking also runs Gemini Live and the voice features in Docs, Gmail, and Keep for paying Google AI subscribers, while the standard model powers Search Live.
