Google Releases Gemini 3.8 Live and Extended Thinking Models

Google launched Gemini 3.8 Live and 3.8 Live Extended Thinking, new state-of-the-art live audio models. The models support 97 languages, near real-time visual understanding, and background tool calling without disrupting conversation.
Key insights
Model Variants and Pricing
Gemini 3.8 Live focuses on high-concurrency, low-latency daily assistant and retrieval services with cost-effective deployment. Gemini 3.8 Live Extended Thinking targets enterprise complex tasks with enhanced multi-step reasoning and planning, featuring 'real-time thinking and speech synchronization' where the model narrates progress.
Key Features
Both models support automatic detection and seamless switching across 97 languages, near real-time multimodal visual understanding (e.g., watching a screen while chatting), and background asynchronous tool calls that allow the model to call APIs or retrieve data without interrupting the voice conversation.
Availability and Use Cases
The models are available in Gemini Live via the Gemini App and through the Gemini API via Google AI Studio. Use cases include programming tutoring, customer service, professional troubleshooting, and multi-system coordination. The Extended Thinking model can narrate task progress to keep the conversation going.