Google has expanded its Gemini 3.8 model lineup with the introduction of Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking. These voice-focused models are designed to facilitate more natural interaction and collaborative work. Gemini 3.8 Live is optimized for high-efficiency tasks, while the Extended Thinking version is built for complex problem-solving.
Gemini 3.8 Live processes visual data in real-time and supports switching between 97 languages during conversations. It manages API calls and external tool interactions in the background without interrupting the user. The model also incorporates SynthID digital watermarking for all generated audio output.
Gemini 3.8 Live Extended Thinking allows for simultaneous reasoning and speech. It currently holds the top position in the Artificial Analysis Speech to Speech Quality Index and the τ-Voice-banking benchmark for real-time fintech customer support tasks.
General users can access Gemini 3.8 Live via Search Live and the Extended Thinking version through Gemini Live. AI Pro and Ultra subscribers gain access to the Extended Thinking model in Google Docs, Gmail, and Keep. Developers can utilize both models via Gemini API and Google AI Studio, while enterprise clients access them through Gemini Enterprise.
