Gemini 3.8 Live Extended Thinking

Gemini 3.8 Live Extended Thinking is the higher-tier variant of Google DeepMind's Gemini 3.8 Live voice model, built for high-complexity tasks. It adds deeper, multi-step reasoning that runs concurrently with speaking: the model can reason while it talks, insert natural verbal acknowledgments ("Let me check that…") and narrate live progress through multi-step tasks. It retains the Live capabilities — near real-time voice and visual processing, 97 languages, tool and API calling, and SynthID audio watermarking.

Released on September 15, 2026. In benchmarks it ranks #1 on Artificial Analysis' Speech-to-Speech Quality Index (82.6), scores 68.6% on τ-Voice, 35.1% on Sierra's τ-Voice-banking and 97.7% on Big Bench Audio. It is available via the Gemini API and Google AI Studio, in Gemini Live and Search Live, and for Google AI Pro/Ultra subscribers in Workspace Docs as well as in Gmail and Keep; enterprise rollouts (Gemini Enterprise) are in preview. Integrators include Salesforce, ServiceNow, Genspark, Lumeris and the Agora, LiveKit and LangChain platforms.