Introducing Gemini 3.8 Live and 3.8 Live Extended Thinking
Google DeepMind introduced Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking, describing them as its most advanced live dialogue models yet, with major upgrades in intelligence and parallel reasoning for near real-time reasoning in voice agents and more intuitive conversation with AI.
Gemini 3.8 Live is built for scale and cost efficiency, combining conversational intelligence with fluid dialogue and visual grounding. Gemini 3.8 Live Extended Thinking is built for high-complexity tasks, offering increased intelligence and multi-step reasoning. Together they provide developers and enterprises the building blocks for reliable, production-ready voice agents, and make speaking with Gemini across the Gemini app, Google Workspace, and Search more fluid and collaborative for complex voice-driven tasks.
Gemini 3.8 Live Extended Thinking delivers enterprise-grade task completion and intelligence, taking the #1 overall spot on Artificial Analysis' Speech to Speech Quality Index (82.6). It leads in agentic task completion with 68.6% on τ-Voice and 35.1% on Sierra's τ-Voice-banking benchmark, and scores 97.7% on Big Bench Audio for reasoning, while maintaining a highly competitive price point versus other frontier models. Gemini 3.8 Live has shown high user preference, securing second place in the Speech Agent Arena, and remains highly cost-effective for scale. On ServiceNow's EVA-Bench, the models push the Pareto Frontier for complex workflows by balancing accuracy with conversational quality (run on the Live API on Gemini Enterprise Agent Platform).
On capabilities, Gemini 3.8 Live processes visual inputs in near real-time to enrich responses, automatically detects and transitions between 97 supported languages mid-conversation, and executes tools and API calls in the background while continuing to talk, so it can acknowledge requests and keep chatting as tasks finish. Gemini 3.8 Live Extended Thinking reasons and speaks simultaneously, delivering increased intelligence for complex workflows without interrupting conversational flow, using early verbal cues like "Let me check that…" to acknowledge prompts.