Gemini 3.8 Live Powers Google Search Live

Google announced the release of its Gemini 3.8 Live and 3.8 Live Extended Thinking models, with Gemini 3.8 Live now integrated into the Google app's Search Live functionality. This rollout commenced on the day of the announcement, enabling real-time conversational AI within Google Search. Rajan Patel, Vice President of Engineering for Search at Google, stated in a written announcement that these new Gemini audio models are powering real-time conversations in Search Live. He elaborated on the benefits, which include more helpful responses that are accompanied by web links for users to explore further information. The feature also provides fluid multilingual support, allowing users to switch languages mid-conversation, and facilitates more natural, free-flowing interactions. To utilize Search Live, users can open the Google app, tap the "Live" icon, and verbally pose their questions. The system will then provide an AI-generated audio response, and users can ask follow-up questions. For users who are visually monitoring the screen, Google will display relevant web links directly on the screen, enabling them to delve deeper into specific topics. Additionally, a "transcript" link is available for users who prefer to review the text version of the conversation, allowing them to see the discussion later. Users can also revisit past Search Live responses by accessing their AI Mode history, picking up conversations where they previously left off. This integration signifies Google's ongoing strategy to embed Gemini models across various Google Search features. Previously, Gemini models have powered features such as AI Overviews, AI Mode, and numerous other components within Google Search. The introduction of Gemini 3.8 Live into Search Live represents a further expansion of this AI integration. Google anticipates that the capabilities and performance of these models will continue to improve, leading to evolving search results over time. The Gemini 3.8 Live Extended Thinking model is also part of this release, although specific details on its immediate application were not provided in the announcement. The core functionality of Search Live involves processing spoken queries and delivering spoken and text-based responses, enhanced by contextual web links and conversational memory.