Google has announced Gemini 3.5 Live Translate for instant text-to-speech translation

Google has been chasing real-time translation for years, which it says is one of its machine learning initiatives. We’ve seen a lot of Google demos in the past, but you’ll need Google phones, headphones, or some other setup. Last year, Google brought real-time translation to many users in the Translations app, and now it’s expanding its availability even further. With Gemini 3.5 Live Translate, you will have the opportunity to translate live in more places and faster than ever before.

The new version of AI is part of the 3.5 family it is set to I/O. Before today, Google only released the Flash version, but we expect the Pro version to drop in the coming weeks. Gemini 3.5 Live Translate is a text-to-speech software that is designed to automatically recognize and translate over 70 languages.

Google says so Gemini 3.5 Live Translate is fast enough for you to communicate effectively, tracking seconds behind the speaker while matching tone, movement, and intonation. In short, voices sound more like you than a typical robot. The demos, all recorded under controlled conditions, sound amazing. You won’t have to wait long to see this model’s prowess, though.

Gemini 3.5 Live Translate is available in several parts of the Google ecosystem. Developers can start building and preview it publicly in Gemini Live API or AI Studio. This template uses language constants and handles all multilingual input tasks automatically, saving developers from having to build manually. It also filters out background noise in busy environments.