Google has introduced Gemini 3.5 Live Translate. The new audio model delivers near real-time speech-to-speech translation in more than 70 languages. It keeps the speaker’s tone, pacing, and pitch, staying just a few seconds behind so conversations sound natural.
The feature is now available to developers via the Gemini Live API and Google AI Studio. For enterprise users, it’s in private preview on Google Meet, while a wider rollout is coming to the Google Translate app on Android and iOS.
Gemini 3.5 Live Translate can automatically handle multilingual inputs and works even in noisy environments. It supports live interpretation for calls, meetings, lessons, and broadcasts. On Android, a new listening mode streams translations directly through the phone’s earpiece, making it easier to use in everyday situations.
Several platforms are already integrating the API, including Agora, Fishjam, LiveKit, Pipecat, and Vision Agents. Grab is testing it to help drivers and riders communicate better in real time. Early feedback points to accuracy, low latency, and reliable language detection.
For Google Meet, translation support expands from 5 languages to 70+, covering over 2,000 language combinations. The private preview starts this month, with broader availability later.
Also Read: Google Try On lets you virtually try on outfits before you buy them
To ensure safety, all AI-generated audio is watermarked with SynthID so that AI-generated content can be identified and flagged.
