Details
- Google for Developers announced Gemini 3.5 Live Translate is now available in public preview via the Gemini API and Google AI Studio.
- The model translates speech as it streams, enabling low-latency, real-time translation suitable for building audio-intensive applications.
- A walkthrough demo shows how developers can use the Gemini Live API to stream audio in and receive translated audio plus text transcription in a single model call.
- Ride-hailing platform Grab has been testing the model to improve communication between drivers and travelers across different languages.
- The model supports multi-language continuous speech translation, including automatic language detection and natural turn-taking during long sessions.
- Gemini 3.5 Live Translate is now available in Google Meet for select Google Workspace customers, with a broader rollout planned later.
- Users can also try the technology in Google Translate using a new listening mode that offers automatic language detection and earpiece audio routing for clearer outdoor listening.
- Google highlights use cases such as digital dubbing tools, conversational translation experiences, and other speech-based products that can be prototyped and built in Google AI Studio.
- Additional technical details, examples, and guidance are provided in an accompanying Google AI blog post linked from the thread.
Impact
By exposing Gemini 3.5 Live Translate through the Gemini API, Google AI Studio, Google Meet, and Google Translate, Google is pushing real-time multilingual communication deeper into mainstream productivity and consumer tools. This strengthens Google’s position against rivals offering live translation and makes speech-to-speech translation more accessible to developers and enterprises.
