Google has officially unveiled Gemini 3.5 Live Translate, its latest AI-powered speech translation technology designed to make multilingual conversations feel more natural. The new model can translate spoken language in near real-time while preserving a speaker’s voice characteristics, helping people communicate across different languages with minimal delay.
The company says Gemini 3.5 Live Translate is now rolling out across several Google products, including Google Translate, Google Meet, Google AI Studio, and the Gemini Live API for developers.
Also read: Google I/O 2026 Highlights Gemini Omni Video AI and Faster Gemini 3.5 Models
Gemini 3.5 Live Translate Aims to Make Conversations More Natural
Traditional translation systems often require users to pause, wait for translations, and then continue speaking. This can make conversations feel slow and less engaging.
Google’s new Gemini 3.5 Live Translate model is designed to solve this problem by processing speech continuously as it is spoken.
According to the company, the AI can:
- Listen to live conversations
- Detect spoken languages automatically
- Generate translated speech in real time
- Preserve tone and speaking style
- Reduce delays between speakers
The result is a smoother conversation experience that feels closer to natural communication.
Supports More Than 70 Languages
One of the biggest highlights of the new translation model is its broad language support.
Google says Gemini 3.5 Live Translate can recognize and switch between more than 70 languages without requiring users to manually select language settings.
This means users can move between different languages more naturally during conversations, making the feature useful for international communication, travel, education, and business meetings.
Near Real-Time Translation With Minimal Delay
The new AI model processes audio as it is being streamed instead of waiting for speakers to finish entire sentences.
As a result, translated speech is delivered only a few seconds behind the original speaker.
Google says this approach helps eliminate many of the awkward pauses commonly associated with traditional translation systems and creates a more fluid conversation experience.
Designed for Multiple Use Cases
Google believes Gemini 3.5 Live Translate can be useful in a wide range of real-world scenarios.
Potential applications include:
International Business Meetings
Teams speaking different languages can communicate more effectively during virtual and in-person meetings.
Live Events and Broadcasts
The technology can help audiences understand content in their preferred language.
Education and Learning
Students and teachers can participate in multilingual lessons more easily.
Customer Support
Businesses can assist customers from different regions without language barriers.
Travel and Everyday Conversations
Travelers can interact with locals more naturally while visiting other countries.
Google Meet Receives a Major Translation Upgrade
Alongside the new model, Google is expanding translation capabilities within Google Meet.
Previously, translation support in Meet was limited to only a handful of languages.
With Gemini 3.5 Live Translate, Google Meet will now support more than 70 languages, significantly expanding its accessibility for global users and organizations.
The upgrade could make multilingual meetings much easier for businesses operating across multiple countries.
Coming to Google Translate on Android and iPhone
Google is also bringing the technology to the Google Translate app.
Users will be able to access Live Translate directly through the mobile application on both Android and iOS devices.
The translated speech can be delivered through compatible connected headphones, allowing users to have conversations without constantly looking at their screens.
New Listening Mode for Android Users
Android users are receiving an additional feature called Listening Mode.
This mode allows translated speech to play directly through the smartphone’s earpiece, creating a more private and convenient translation experience.
The feature could be particularly useful during travel, meetings, or one-on-one conversations.
Works in Challenging Environments
Real-world conversations often take place in noisy settings where background sounds can interfere with speech recognition.
Google says Gemini 3.5 Live Translate has been designed to remain effective even in unpredictable and noisy environments.
This could improve translation accuracy in places such as:
- Airports
- Conferences
- Public spaces
- Restaurants
- Busy workplaces
AI-Generated Audio Includes SynthID Watermarking
As concerns around AI-generated content continue to grow, Google is incorporating its SynthID watermarking technology into generated audio.
According to the company, every audio output created by Gemini 3.5 Live Translate includes an embedded watermark.
This invisible identifier is designed to help distinguish AI-generated content from human-created audio and improve transparency.
Developers Can Access the Technology
Google is also opening access to developers through the Gemini Live API and Google AI Studio.
Developers can use the technology to build custom translation applications and voice-based experiences.
Google highlighted compatibility with several platforms and development tools to simplify integration and deployment.
Global Rollout Begins
Gemini 3.5 Live Translate is now beginning its global rollout across supported Google services.
As availability expands, users and developers worldwide will gain access to one of Google’s most advanced speech translation technologies to date.
Also read: Gemini iOS App Spotted With New Animated UI and Simplified Layout
Final Thoughts
With Gemini 3.5 Live Translate, Google is taking another major step toward breaking down language barriers through artificial intelligence. By supporting more than 70 languages, preserving natural speech characteristics, and delivering translations in near real-time, the company aims to make multilingual conversations more seamless than ever before.
Whether used for travel, education, business meetings, or everyday communication, the new translation model could significantly improve how people connect across different languages around the world.