Google announced today the launch of Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking, calling them its "most advanced real-time conversational models." Following the March release of 3.1 Flash Live, these new models are designed to make conversations with AI more intuitive and intelligent. As the name suggests, Gemini 3.8 Live Extended Thinking offers "enhanced intelligence and multi-step reasoning" suited to highly complex tasks, such as the recently launched Gmail Live (conversational search), Docs Live (draft generation and editing) and Keep Live (note creation).
The model is also rolling out on Gemini Live, where 3.8 Live Extended Thinking can "reason and speak simultaneously."
It delivers enhanced intelligence for complex workflows while maintaining a coherent conversational flow, using early verbal cues such as "let me check…" to acknowledge prompts naturally, and guides users step by step through multi-step background tasks with real-time progress narration. On benchmarks, it ranks "first overall (82.6) on the human-evaluated Speech-to-Speech Quality Index"; "leads on agentic task completion rate with a score of 68.6% over τ-Voice"; "scores 35.1% on Sierra's τ-Voice-banking benchmark"; and "demonstrates strong reasoning capabilities, scoring 97.7% on Big Bench Audio, while remaining highly competitively priced against other frontier models."
Gemini 3.8 Live delivers enhanced conversational intelligence
Meanwhile, the base Gemini 3.8 Live model offers "conversational intelligence with fluid dialogue and visual grounding." It can process "visual input" near-instantly and run tools and API calls in the background while a conversation continues. Users can expect it to "respond to requests and keep the conversation going" as background tasks complete. It can also detect and switch between 97 supported languages mid-conversation. The model is "built for scale and cost efficiency" and "ranks second on the Voice Agent Arena."
Gemini 3.8 Live will power the Search Live experience in AI Mode.