On June 9, 2026 (Beijing Time), Google proudly announced the release of its latest innovation in real-time audio translation: the Gemini 3.5 Live Translate model. This groundbreaking model facilitates seamless voice-to-voice translation in over 70 languages, delivering translations that are not only natural and fluent but also retain the original speaker's unique tone, pace, and pitch. Unlike traditional translation systems that operate on a turn-by-turn basis, Gemini 3.5 Live Translate optimally balances the need for contextual information to enhance translation accuracy with the demand for instant responses. This balance ensures a smooth and uninterrupted communication flow, with delays kept to a minimum of just a few seconds.
From today onwards, developers can delve into the public preview version of this model via the Gemini Live API and Google AI Studio. Enterprise users are invited to explore a private preview within the Google Meet environment. Meanwhile, the general public can access this service through the Google Translate app on both Android and iOS platforms. Furthermore, a dedicated 'Listening Mode' has been introduced for Android devices, enabling users to conveniently listen to translation results directly through their phone's earpiece.