OFICIAL The Keyword

Introducing Gemini 3.8 Live and 3.8 Live Extended Thinking

What happened
Based on The Keyword · Sep 15, 2026

Google introduced two new AI models, Gemini 3.8 Live and 3.8 Live Extended Thinking, designed to enhance real-time voice interactions with improved reasoning and task execution capabilities.

Introducing Gemini 3.8 Live and 3.8 Live Extended Thinking
The Keyword — Google
Key points
·
Gemini 3.8 Live Extended Thinking leads in agentic task completion with 68.6% on τ-Voice benchmark
·
Gemini 3.8 Live processes visual inputs in near real-time and supports 97 languages mid-conversation
·
Gemini 3.8 Live Extended Thinking uses verbal cues like “Let me check that…” to acknowledge prompts
Key numbers
·
6% on τ-Voice and 35.
·
1% on Sierra’s τ-Voice-banking benchmark.
·
7% on Big Bench Audio while maintaining competitive pricing compared to other frontier models.

Google announced the launch of Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking, positioning them as the company’s most advanced live dialogue models to date. These upgrades focus on enhancing parallel reasoning and intuitive collaboration, enabling users to execute complex tasks using voice commands without interruption. The models are now accessible via the Gemini API, Google Workspace, and the Gemini app, supporting real-time visual context and background task execution.

Gemini 3.8 Live Extended Thinking is tailored for enterprise-grade task completion, achieving a score of 82.6 on Artificial Analysis' Speech to Speech Quality Index and leading in agentic task completion benchmarks with 68.6% on τ-Voice and 35.1% on Sierra’s τ-Voice-banking benchmark. It also scored 97.7% on Big Bench Audio while maintaining competitive pricing compared to other frontier models. The model prioritizes cost-effectiveness and scalability for developers and enterprises.

Gemini 3.8 Live processes visual inputs in near real-time, enriching conversations with contextual responses while automatically detecting and transitioning between 97 supported languages mid-conversation. It executes background tasks such as API calls without disrupting the dialogue, allowing users to continue chatting while tasks complete. The model demonstrates capabilities like guiding employee onboarding in real time and playing chess using visual context and natural conversational flow.

Gemini 3.8 Live Extended Thinking integrates deeper reasoning with uninterrupted conversational flow, using verbal cues like “Let me check that…” to acknowledge prompts and narrating progress for multi-step tasks. It supports complex workflows such as transforming sketches into functional React components and coordinating multi-step bookings without interrupting speech. The models are available in Google Workspace via Docs Live, Gmail Live, and Keep Live, and developers can access them through the Gemini Live API for building voice-driven interfaces.

Original source → Deals on Clipraptor.com →