Google DeepMind Unveils Gemini 3.8 Live Models With Extended Thinking

15 Sept 2026, 10:59 pm · 17h ago · 1 min read · OfficeChai

Google DeepMind has introduced two new real-time conversational audio models, Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking. The releases aim to compete against frontier voice models from OpenAI and Anthropic by offering low-latency multimodal reasoning at competitive operational price points. The Extended Thinking model incorporates deeper chain-of-thought reasoning during real-time speech interactions, allowing the system to tackle complex multi-step user queries dynamically without interrupting conversational flow across developer tools, customer service platforms, and enterprise virtual assistant integrations.