Natural live dialogue
The model is designed for fast, fluid voice interaction and is positioned for scalable, cost-efficient voice-agent workloads.
Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking
Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking are Google voice-dialogue models for near-real-time conversation, visual context, tool use, and reasoning. They support developers and enterprises building voice agents, as well as live experiences across Gemini, Workspace, and Search.
Overview
Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking are Google models for near-real-time voice interaction. They combine conversational dialogue with visual grounding, tool use, and reasoning so users can work through tasks by speaking rather than relying only on typed prompts.
The standard Live model is aimed at fluid, scalable voice agents and fast conversational responses. Live Extended Thinking is intended for more demanding workflows that require multi-step reasoning; it can continue speaking while work proceeds and narrate progress as background tasks advance. The models are presented for developer and enterprise voice applications, as well as experiences in the Gemini app, Google Workspace, and Search.
Core capabilities
Near-real-time visual grounding
It accepts visual context in near real time, allowing conversations to incorporate what the model is seeing, such as an onboarding screen or a chessboard.
Multilingual switching
Gemini 3.8 Live can detect and transition between 97 supported languages during a conversation.
Background task execution
The models can execute tools and API calls in the background while acknowledging requests and continuing the conversation.
Extended reasoning with live progress
The Extended Thinking variant is designed for high-complexity tasks, using multi-step reasoning while speaking and providing early verbal cues and progress narration.
Voice-agent development ecosystem
The Gemini Live API supports voice-agent development, while platforms including Agora, Fishjam, LangChain, LiveKit, Pipecat, Vercel, and Vision Agents are cited as enabling development and deployment workflows.
Practical use cases
Conversational task automation
A voice agent can acknowledge a request, call APIs or tools in the background, and keep the user informed without forcing a silent wait. This suits workflows such as bookings and other multi-step service interactions.
Visual workplace assistance
Employees can receive live guidance during onboarding or troubleshooting while the model interprets the screen or other visual context in near real time.
Production voice agents
Developers can build voice-driven interfaces using the Gemini Live API and supporting real-time media platforms, focusing on the interaction design while those platforms handle streaming infrastructure.
Complex creation and planning
Extended Thinking can work through complex plans, explain progress aloud, and turn spoken instructions into outputs such as functional React components or business planning materials.
Multilingual conversations
Multilingual conversations can move between supported languages without requiring a manual mode change, which is useful for live dialogue involving language changes.
Pros and Cons
Pros
- Supports natural, near-real-time voice dialogue with visual context.
- Can run tools and API calls in the background without stopping the conversation.
- Extended Thinking adds multi-step reasoning, verbal acknowledgment, and progress narration for complex tasks.
- Supports automatic switching across 97 languages during a conversation.
- Provides a documented path for developers through the Gemini Live API and Google AI Studio.
Cons
- The supplied launch material does not provide official pricing, plan limits, setup requirements, or complete availability details.
- The article describes benchmark results and selected product experiences, but does not establish performance for every voice workflow or deployment environment.
- Access differs by surface: the source describes developer access and enterprise rollout separately, so availability should be checked for the intended platform.
FAQ
Where can developers access Gemini 3.8 Live?
Gemini 3.8 Live is available through the Gemini API and Google AI Studio for developers. The source also says it is rolling out in Gemini Enterprise, with further enterprise availability information incomplete in the supplied text.
What is the difference between Gemini 3.8 Live and Live Extended Thinking?
Gemini 3.8 Live is intended for fast, fluid dialogue, visual grounding, and scalable voice-agent workloads. Gemini 3.8 Live Extended Thinking is intended for more complex tasks that benefit from multi-step reasoning and deeper task completion.
Can the models continue a conversation while working on a task?
The models can process visual input in near real time, continue speaking while tools and API calls run in the background, and, in Extended Thinking, provide verbal progress cues during longer tasks.
How many languages can Gemini 3.8 Live handle?
Gemini 3.8 Live automatically detects and transitions between 97 supported languages during a conversation, according to the launch article.
Is generated audio watermarked?
Audio generated by Google AI products is watermarked with SynthID. The source describes this as an imperceptible watermark intended to help detect AI-generated audio.
Quick Facts
- Category
- AI voice and dialogue models
- Variants
- Gemini 3.8 Live; Gemini 3.8 Live Extended Thinking
- Primary users
- Developers, enterprises, and users of Gemini app, Workspace, and Search experiences
- Developer access
- Gemini API and Google AI Studio
- Core workflow
- Live conversation with visual input, background tools, and API calls
- Audio provenance
- Google says generated audio is watermarked with SynthID
Alternatives à Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking
Flunkey
Flunkey is a voice-first productivity layer for Windows that converts speech into text, AI-assisted outputs, and remembered context. It is built for users who want to capture ideas and actions without leaving the app they are working in.
PXZ AI
Une plateforme IA tout-en-un qui combine des outils pour l'image, la vidéo, la voix, l'écriture et le chat afin d'améliorer la créativité et la collaboration.
Gemma AI
Gemma AI is a phone call reminder app that calls you with scheduled reminders instead of push notifications. It helps people who want a more direct way to stay on schedule, with Google Calendar sync and conversational call interactions.
Grok 4.7
Grok 4.7 is an xAI model for software development and knowledge work, including extended coding tasks and professional analysis. It is available through Grok Build, Cursor, the Grok API, and other developer platforms, with token pricing starting at $2 per million input tokens and $6 per million output tokens.
Wallie
Wallie is an open-source AI streamer that watches your screen, hears chat, and generates live commentary in a configurable persona. It runs locally on your machine with your own keys and is aimed at faceless content, autonomous streams, and real-time reactions.
Claude Overlay
Claude Overlay is a Windows desktop overlay for Claude Code that reads your screen so you can ask questions, inspect content, and request edits without leaving the current app. It runs on your existing Claude subscription through the Claude CLI, so no separate API key is required.