AIGoogle Releases Gemini 3.8 Live with Extended Thinking for Real-Time Voice AI
Google launched Gemini 3.8 Live models optimized for cost-efficient, real-time voice interactions with an Extended Thinking variant for simultaneous reasoning and speaking. The models topped relevant speech-to-speech quality benchmarks and support advanced agentic voice applications with background tool use.
TLDR
Advances in multimodal and real-time AI capabilities strengthen Google's competitive position in voice and agent markets. Benchmark improvements and developer-friendly features signal accelerating capability deployment, contributing to the broader industry race for frontier model advantages.
Combined views
—
2 Sources, first seen 1d ago
Google Releases Gemini 3.8 Live with Extended Thinking for Real-Time Voice AI
Google launched Gemini 3.8 Live models optimized for cost-efficient, real-time voice interactions with an Extended Thinking variant for simultaneous reasoning and speaking. The models topped relevant speech-to-speech quality benchmarks and support advanced agentic voice applications with background tool use.
TLDR
Advances in multimodal and real-time AI capabilities strengthen Google's competitive position in voice and agent markets. Benchmark improvements and developer-friendly features signal accelerating capability deployment, contributing to the broader industry race for frontier model advantages.