TechGoogle releases Gemini 3.8 Live and Extended Thinking for real-time voice AI
Google launched speech-to-speech models in Gemini API, Studio, Search, and Workspace. Gemini 3.8 Live offers low-latency conversational flow with interruption support in 97 languages. Extended Thinking variant adds background reasoning and multi-step problem-solving while speaking.
TLDR
These releases raise the bar for voice agents and real-time multimodal AI capabilities, making advanced features more accessible. The models top benchmarks like the Speech to Speech Quality Index, signaling commoditization of sophisticated voice AI across Google's product ecosystem.
Combined views
8
1 post, first seen 22h ago
Google releases Gemini 3.8 Live and Extended Thinking for real-time voice AI
Google launched speech-to-speech models in Gemini API, Studio, Search, and Workspace. Gemini 3.8 Live offers low-latency conversational flow with interruption support in 97 languages. Extended Thinking variant adds background reasoning and multi-step problem-solving while speaking.
TLDR
These releases raise the bar for voice agents and real-time multimodal AI capabilities, making advanced features more accessible. The models top benchmarks like the Speech to Speech Quality Index, signaling commoditization of sophisticated voice AI across Google's product ecosystem.