Google launches Gemini 3.8 Live and Live Extended Thinking voice models
A post citing Google describes automatic detection for 97 languages, near-real-time visual understanding and tools that run in the background while conversations continue.
TLDR
A September 17 post citing Google said Extended Thinking led Artificial Analysis’ Speech to Speech Quality Index at 82.6. It described the models as rolling out through the Gemini API, Google AI Studio, Search Live, Gemini Live and Workspace for eligible subscribers. A separate post that day listed input-audio prices of $0.84 per hour for standard Gemini 3.8 Live and $3.50 for Extended Thinking, compared with $5.83 for GPT-Live-1 (Astra).
Combined views
—
2 Sources, first seen ago
