Google Launches Gemini 3.8 Live Voice Models

Google introduced Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking, which it calls its most advanced live dialogue models yet, aimed at voice agents and at spoken work across the Gemini app, Google Workspace and Search. Google says Extended Thinking takes the top overall spot on the Artificial Analysis Speech to Speech Quality Index with a score of 82.6 and leads agentic task completion with 68.6% on τ-Voice and 35.1% on Sierra's τ-Voice-banking benchmark, while Gemini 3.8 Live placed second in the Speech Agent Arena.

Gemini 3.8 Live processes voice and visual input, switches between 97 supported languages during conversations, and runs tools or API calls while continuing to speak. Extended Thinking supports more complex tasks by reasoning and speaking simultaneously, providing verbal updates as it completes multistep work. Google says all audio generated by its AI products carries its SynthID watermark. Developer platforms including Agora, LangChain, LiveKit, Pipecat and Vercel support the Gemini Live API, and Google names Salesforce, Genspark and Lumeris among partners.

Both models are rolling out through the Gemini API and Google AI Studio. Gemini 3.8 Live is also available in Search Live, while Extended Thinking is reaching Gemini Live and eligible Google AI subscribers in Docs, Gmail, and Keep. Enterprise versions are available in private preview, with wider access planned for Gemini Enterprise, customer service products, and Google Workspace business customers.