Twenty years ago, Google Translate began as one of our first experiments in machine learning to turn the science of language into the magic of human communication. That test has come a long way with more than three billion words being translated to billions of users across our products every month.
Today, we're taking our next step with the release of Gemini 3.5 Live Translate, our latest audio model for live speech-to-speech translation.
The model automatically detects 70+ languages and produces smooth, natural-sounding translated speech that preserves the speaker's pronunciation, movement and intonation. Unlike turn-by-turn systems that wait for the speaker to finish speaking before responding, 3.5 Live Translate generates speech continuously, balancing the trade-off between waiting for context to improve quality and translating quickly to always match the speaker. It delivers fluid sound without lag and stays just a few seconds behind the speaker during a session.
Gemini 3.5 Live Translate is rolling out starting today for all Google products:
Build with 3.5 Live Translate
Gemini 3.5 Live Translate processes speech as it is broadcast, allowing for more seamless communication across languages. The model handles multilingual input without the need to configure settings manually. At the same time, its sound capabilities ensure that applications can handle noisy, unpredictable environments. You can use its capabilities to help facilitate live translation of multilingual calls, meetings, courses, broadcasts and more.
