$GOOGL DeepMind Launches Gemini 3.8 Live and Extended Thinking Models With Real-Time Speech and Visual...

Google DeepMind introduced Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking, advanced conversational AI models supporting real-time speech, visual understanding across 97 languages, and background tool use without interrupting conversations.
Extended Thinking is designed for more complex tasks, reasoning through problems while narrating its progress in real time, including coding and interactive visual work.
Google says the models lead several benchmarks, including 82.6% on the Speech-to-Speech Index and 68.6% on agentic performance, while offering audio input at $0.84 per hour.