Google Launches Gemini 3.8 Live and Extended Thinking Models
September 15, 2026
Gemini 3.8 Live and 3.8 Live Extended Thinking are now available via the Gemini API for real-time voice dialogue. The models support parallel reasoning, real-time visual context, and background task execution during live interactions.
HOW THIS AFFECTS YOU
●
builderYou can now integrate low-latency, multimodal voice reasoning directly into your applications via API.
●
designerThis enables new UI patterns for natural, non-interruptive voice-based human-AI collaboration.