Natural voice translation with Gemini 3.5 Live Translate

Gemini 3.5 Live Translate text and star icon on a light blue background.

Upgraded, near real-time voice translation

Twenty years ago, translation at Google began as one of our pioneering machine learning experiments. This week we took our next big leap with Gemini 3.5 Live Translate. This new audio model for live speech-to-speech translation automatically detects more than 70 languages while preserving the speaker's natural intonation and pacing, eliminating awkward pauses. You can now have fluid, near-real-time conversations during multilingual calls, meetings or travel, removing language barriers in seconds. It's rolling out across Google Translate, Gemini Live API and Google AI Studio.

Learn more

Popular on the Keyword

'Create a...' chat bubble surrounded by content creation icons.

NOTEBOOKLM

Do better research with NotebookLM

SAFETY & SECURITY

How we're combatting AI scams with security, legislation and more

Paris Hilton holding a Chihuahua and smartphone in front of a Google building.

ANDROID

Paris Hilton is Android’s first icon in residence

GOOGLE FI

Get travel-ready with Google Fi Wireless

GEMINI APP

Save time and grow your business with new Gemini tools

Man in a white hard hat and yellow safety vest with arms crossed.

GOOGLE.ORG

Growing the next generation of American workers

A dense pile of stylized basketballs in shades of orange and light teal.

What's trending on Search during the 2026 NBA Finals

Beyond the news

Gemma is our family of open AI models, and Gemma 4 is our most capable version yet — allowing developers to run it on their own hardware to build great apps and experiences. In fact, the Gemma 4 model family has officially crossed 150 million downloads. With so many people building, we wanted to highlight a few projects that caught our eye. See how three builders are using Gemma 4 to push creative boundaries and create entirely new experiences.