Google Launches Gemini 3.5 Live Translate: Real-Time Voice Translation Comes to Meet & Translate

Google today officially launched Gemini 3.5 Live Translate, its latest voice-to-voice real-time translation model. Unlike traditional “turn-based” conversational translation, 3.5 Live Translate generates continuous speech, striking a balance between waiting for context to improve quality and delivering instant translation for smooth, synchronized output. The model is now rolling out to the Google Translate app on Android and iOS, supporting fluent, natural voice translation in over 70 languages.

For Android users, Google Translate also introduces a new “Listen Mode”, allowing users to hear translated results directly through the phone’s earpiece without needing headphones. Additionally, Gemini 3.5 Live Translate is being integrated into Google Meet, enabling real-time cross-language communication in video meetings. On the developer side, the model is available in public preview via the Gemini Live API and Google AI Studio, making it possible to integrate this capability into third-party apps.

This update marks a major step forward for Google in real-time translation—shifting from intermittent, turn-based responses to coherent, low-latency speech streams—and has the potential to significantly improve cross-border communication and collaboration. With the developer ecosystem now open, more application scenarios built on this model are likely to emerge in the near future.