Gemini Audio
Live dialogue
Fluid and natural live dialogue and translation capabilities, for powerful voice-first applications
Models
Create agents capable of handling complex tasks and using tools, while engaging in natural conversations.
Slide 1 of 3
3.8 Live Extended Thinking
Best for complex reasoning and high-complexity tasks. Thinks and narrates its progress in real time to solve difficult, multi-step challenges.
3.8 Live
Best for efficient high-volume deployments. Delivers levels of reasoning required for custom voice agents while keeping costs low.
3.5 Live Translate
Best for near real-time speech-to-speech translation. Overcomes language barriers across 70+ languages while maintaining the speaker’s natural tone and rhythm.
Capabilities
Build natural-sounding and reliable voice agents with our live dialogue models.
Slide 1 of 4
Multi-tasking
Handles background tasks while speaking, using deep reasoning to solve complex queries on the fly – without stalling the conversation.
Visual Understanding
Understands visual input to grasp deeper context during conversations. This allows for richer, more natural, and more relevant responses.
Proactive listening
Knows the difference between direct engagement and background chatter. Understands the rhythm of speech, knowing precisely when to talk – and when to stay silent.
Alphanumeric accuracy
Handles complex alphanumeric data – like phone numbers, order IDs, and email addresses– reliably and accurately.
Build voice agents
See how you can use 3.8 Live Extended Thinking to build voice agents.
Live speech translation
Overcomes language barriers by using Gemini’s speech-to-speech translation capabilities.
Broad language coverage
Delivers fluid speech-to-speech translation across 70+ languages and 2,000 language pairs.
Consistent intonation
Preserves the speaker’s original intonation, pacing and pitch to capture not just what they said, but how they said it.
Multilingual input
Translates multiple languages in a single session – no need to change the settings.
Automatic language detection
Identifies the language being spoken and begins translation, without being told what it is.
Low latency
Minimizes processing lag to eliminate awkward pauses – and keep conversations flowing naturally.
Showcase
Slide 1 of 12
“At 11Sight, we build AI voice agents that handle real customer conversations at scale. Integrating Google's Gemini 3.8 Live has meaningfully elevated the naturalness and responsiveness of our agents. We deliver faster, more fluid interactions across service, sales, and front desk operations. This translates directly into higher call resolution rates and better customer experiences.”
Dr Farokh Eskafi
Co-Founder & CTO, 11Sight
“We're excited that Gemini 3.8 Live and Agentforce are coming together to reimagine what's possible in intelligent service. By bringing together real-time, multimodal capabilities with agentic AI, we're exploring new ways to create richer, more intuitive customer experiences from first contact to resolution.”
Bob Van Osten
VP of Product for Agentforce, Salesforce
“These models feel genuinely differentiated rather than like minor variations of the same system. Gemini 3.8 Live delivers the speed needed for highly conversational experiences, while Gemini 3.8 Live Extended Thinking brings stronger reasoning for complex, production-grade voice agents. Together, these models feel like a real toolbox for voice AI, with speed when you need it and deeper reasoning when the workflow demands it.”
Mason Adams
Developer Evangelist, Agora
“Ambr AI trains people for high-stakes workplace conversations, so our AI characters have to hold up under pressure — objection handling, negotiation maths, sudden language switches. With Gemini 3.8 Live Extended Thinking, our personas can now work through multi-step pricing arithmetic without pauses. That's the difference between a training exercise and a conversation that feels real.”
Jamie Wood
Co-founder, Ambr AI
“In medicine, training international staff on realistic patient interactions makes a real difference. Using Gemini 3.8 Live Extended Thinking, our learners get contextual support while the agent creates the next practice scenario in the background. The speed and agentic capabilities make medical communication training feel more natural, responsive, and effective.”
Dr. med. Rüdiger Schmitz
CEO, casuu
“We're building a personal AI that knows you, speaks your language, and is always on your side. Today, it handles over a million live calls daily across nine Indian languages. Gemini 3.8 Live improved latency, interruption handling, multilingual conversations, and tool-call reliability. This AI doesn't just answer calls; it gets things done for you.”
Akhilesh Damaraju
CEO, Equal AI
“Genspark builds voice-driven assistants that act for users in the real world. We tested Gemini 3.8 Live in our most demanding scenarios: real telephony audio, live tool calling mid-conversation, and continuous video understanding. Across hundreds of live sessions, it delivered natural turn-taking, reliable barge-in, and clean function calling. We're excited about what it unlocks for real-time voice agents.”
Kay Zhu
Co-founder & CTO, Genspark
“With B by Lenskart, we set out to build an everyday AI companion for users, and the Gemini Live model has been central to realizing that vision. Its natural conversations, robust tool calling, and ability to understand emotion, tone, and context across India's many languages and dialects have helped us create a truly intuitive experience.”
Kushal Agarwal
Business Head Smartglasses, Lenskart
“Gemini 3.8 Live and 3.8 Live Thinking both vitalize conversations with realism and natural voice inflections. The depth of responses is great for covering a multitude of use cases.”
Tim Nguyen
Software Engineer, LiveKit
“Gemini 3.8 Live brings crucial improvements to clinical voice tech and delivers a much more natural, controlled pace. It follows instructions so tightly that we can finally cut out the over-prompting, leaving us with reliable tool execution, natural interruption handling, and just a tiny, charming touch of upbeat vocal inflections.”
Anagha Vyas
Vice President of Artificial Intelligence, Lumeris
“We’ve always been focused on building user-first products that address real-world problems, and we’re excited to partner with Google DeepMind to explore novel ways to overcome language barriers between our passengers and driver partners. Testing Gemini 3.5 Live Translate, we’ve been impressed by its ability to auto-detect multiple languages, and translate speech accurately with low latency.”
Philipp Kandal
Chief Product Officer, Grab
“Gemini 3.5 Live Translate paired with Fishjam’s MoQ protocol sets a new frontier for near real-time multimedia streaming, allowing speech-to-speech translation into 90 languages.”
Maciej Rys
VP of Engineering, Software Mansion
Performance
Gemini 3.8 Live & 3.8 Live Extended Thinking introduces major leaps over our predecessor, Gemini 3.1
Slide 1 of 5
Gemini 3.8 Live Extended Thinking provides enterprise-grade task completion and intelligence, capturing the #1 overall spot on Artificial Analysis' Quality Index (82.6%).
Gemini 3.8 Live Extended Thinking leads in agentic task completion with 68.6% on τ-Voice.
Gemini 3.8 Live Extended Thinking leads in agentic task completion with 35.1% on Sierra’s τ-Voice-banking benchmark.
Gemini 3.8 Live has shown a high preference among users, securing a second place in the Speech Agent Arena. In addition to this performance, it remains highly cost-effective—providing developers and enterprises with a capable and efficient model built for scale.
On ServiceNow’s EVA-Bench, a benchmark for evaluating voice agents, our models push the Pareto Frontier for complex workflows by successfully balancing accuracy with conversational quality.
Model information
| Name | 3.8 Live Extended Thinking | 3.8 Live | 3.5 Live Translate |
|---|---|---|---|
| Status | General Availability | General Availability | General Availability |
| Input |
|
|
|
| Output |
|
|
|
| Input tokens | 128k | 128k | 128k |
| Output tokens | 64k | 64k | 64k |
| Availability |
|
|
|
| Documentation | View developer docs | View developer docs | View developer docs |
| Model card | View model card | View model card | View model card |
Try Live dialogue
Google AI Studio
The fastest path from prompt to production
Gemini API
Get started with cutting-edge AI models
Gemini Live API
Low-latency, real-time voice and video interactions with Gemini
Gemini Enterprise Agent Platform
Build, scale, and govern agents
Gemini Enterprise for Customer Experience
Deploy specialized agents for product discovery, shopping, and customer service
Google Translate
Understand your world and communicate across languages
Google Workspace
Collaborate, create, and communicate all in one place