Gemini 3.8 Live and 3.8 Live Extended Thinking

Google has unveiled two new AI models, Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking, designed to enhance voice interactions. These models offer improved naturalness, fluidity, and intelligence in conversation, capable of handling interruptions, multiple languages, and explaining their thought processes. They aim to make AI feel like a real conversation partner, a significant step towards more intuitive AI interactions.

These models are built for developers and enterprises, providing the tools to create reliable, production-ready voice agents. They currently lead in agentic task completion, with a 68.6% score on the τ-Voice benchmark, and offer strong reasoning capabilities, scoring 97.7% on Big Bench Audio. Notably, they maintain competitive pricing compared to other frontier models.

Gemini 3.8 Live, in particular, has gained favor among users, ranking second in the Speech Agent Arena. It also processes visual inputs in near real-time, automatically detecting and switching between 97 supported languages mid-conversation. It can execute tools and API calls in the background while continuing the conversation. When deeper reasoning is required, Gemini 3.8 Live Extended Thinking can reason and speak simultaneously, acknowledging prompts naturally and providing live progress narration for multi-step background tasks.

These advancements are integrated across Google Workspace and Search, delivering more intuitive, collaborative experiences, especially for complex tasks. The Gemini Live API enables developers to build high-performance voice-driven interfaces using platforms like Agora, Fishjam, LangChain, LiveKit, Pipecat, Vercel, and Vision Agents. All audio generated by Google's AI products is watermarked with SynthID, ensuring detectability to prevent misinformation.