Google has taken another leap forward in the world of translation with the introduction of Gemini 3.5 Live Translate. This cutting-edge audio model is set to transform live speech-to-speech translation, offering a seamless experience for users worldwide. With the ability to detect over 70 languages, Gemini 3.5 generates smooth, natural-sounding translations that mirror the speaker's intonation and pitch. Unlike traditional systems, it provides continuous translation, maintaining a fluid audio stream just seconds behind the speaker.
Rolling out across various Google platforms, Gemini 3.5 is available to developers via the Gemini Live API and Google AI Studio, and is also being tested by enterprises in Google Meet. For everyone else, it’s accessible through the Google Translate app on both Android and iOS. The model’s ability to handle multilingual inputs without manual configuration makes it a game-changer for multilingual calls, meetings, and broadcasts.
Developers are already integrating Gemini 3.5 into platforms like Agora and LiveKit, simplifying the process of building voice translation apps. Grab, a partner in this initiative, is testing the model to enhance communication between drivers and travellers, making over 10 million voice calls monthly.
Gemini 3.5 Live Translate is also set to enhance Google Meet, expanding language offerings and improving interface accessibility. With its rollout on Google Translate, users can experience seamless translations through headphones or a new ‘listening mode’ that streams translations directly to the phone’s earpiece.
All audio generated by Gemini 3.5 is watermarked with SynthID, ensuring content authenticity and preventing misinformation. This innovation marks a significant milestone in Google's journey to connect the world through language.
Want to hear more? Join Mal & Matt on the Property AI Report Podcast each week!
Access from your preferred podcast provider by clicking here
Made with TRUST_AI - see the Charter: https://www.modelprop.co.uk/trust-ai