Gemini 3.8 Live Extended Thinking powers Gemini Live, Gmail, & Keep

Google announced Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking, which it calls its "most advanced live dialogue models yet," following the March launch of Gemini 3.1 Flash Live. Gemini 3.8 Live Extended Thinking scored 82.6 for the #1 overall spot on Artificial Analysis' Speech to Speech Quality Index, 68.6% on τ-Voice for agentic task completion, 35.1% on Sierra's τ-Voice-banking benchmark, and 97.7% on Big Bench Audio, and is rolling out to Gemini Live today to power Gmail Live, Docs Live, and Keep Live. The base Gemini 3.8 Live model handles near real-time visual inputs, switches between 97 supported languages mid-conversation, took second place in the Speech Agent Arena, and will power AI Mode's Search Live experience.

Google today announced https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-8-live-gemini-3-8-live-extended-thinking/ Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking as its “most advanced live dialogue models yet.” Following the launch of 3.1 Flash Live in March https://9to5google.com/2026/03/26/gemini-3-1-flash-live/ , the new models aim to “make conversing with AI feel more intuitive and intelligent.” As the name suggests, Gemini 3.8 Live Extended Thinking offers “increased intelligence and multi-step reasoning.” It’s meant for high-complexity tasks like the recently launched Gmail Live https://9to5google.com/2026/09/03/gmail-docs-keep-live/ conversational search , Docs Live draft generation and editing , and Keep Live note creation . The model is also beginning to roll out to Gemini Live today. 3.8 Live Extended Thinking “reasons and speaks simultaneously.” It delivers increased intelligence for complex workflows while maintaining an uninterrupted conversational flow — using early verbal cues like “Let me check that…” to acknowledge prompts naturally, and live progress narration to walk users through multi-step background tasks as they progress. In terms of benchmarks: - “…the 1 overall spot on Artificial Analysis’ Speech to Speech Quality Index 82.6 ” - “…leads in agentic task completion with 68.6% on τ-Voice” - “…35.1% on Sierra’s τ-Voice-banking benchmark” - “…strong reasoning capabilities, scoring 97.7% on Big Bench Audio, while maintaining a highly competitive price point compared to other frontier models.” Meanwhile, the Gemini 3.8 Live base model offers “conversational intelligence with fluid dialogue and visual grounding.” The model can process “visual inputs in near real-time,” with tools and API calls executing in the background while the conversation continues. Users can expect the model to “acknowledge requests and keep chatting while tasks finish in the background.” Additionally, it can detect and transition between 97 supported languages mid-conversation. It’s “built for scale and cost efficiency” with a “second place in the Speech Agent Arena https://artificialanalysis.ai/speech-to-speech?api-benchmarks=agentic-performance-vs-cost-to-run speech-to-speech-arena-results-tabs .” Gemini 3.8 Live will power AI Mode’s Search Live experience. FTC: We use income earning auto affiliate links. More. https://9to5mac.com/about/ affiliate our homepage http://9to5google.com/ for all the latest news, and follow 9to5Google on exclusive stories https://9to5google.com/feature/exclusive/ , reviews https://9to5google.com/guides/reviews/ , how-tos https://9to5google.com/guides/how-to/ , and subscribe to our YouTube channel https://www.youtube.com/9to5google