Gemini 3.5 Live Translate becomes your personal interpreter across 70+ supported languages
Google Translate introduced an AI coach for pronunciation practice on Android earlier in April this year. Now the search giant has announced Gemini 3.5 Live Translate, a new AI-powered audio model that aims to make conversations across different languages feel far more natural. Unlike older translation tools that waited for you to finish a sentence before responding, this one keeps pace with you as you speak, carrying over your turn, pacing, and pitch in the output.
Why Gemini 3.5 Live Translate is a big deal
Gemini 3.5 Live Translate can automatically detect the language being spoken without requiring users to manually switch settings. It is designed to start translating almost immediately when the person is speaking, allowing conversations to flow more smoothly as it processes audio continuously. The translated speech is naturally sounding rather than robotic.
The feature is now rolling out inside the Google Translate app on Android and iOS for everybody across the world. If you just want it for travel or day-to-day use, it’s already on your phone.
The AI-powered audio model also benefits Google Meet significantly. The speech translation expands from just five supported languages to more than 70, and the upgrade also enables over 2,000 language combinations in a single meeting, moving beyond the previous limitation of translating mainly to and from English. This is a complete overhaul instead of just a minor update.
If you use Google Meet for international calls, this will directly improve your experience.
However, at the moment, the feature is rolling out in private preview for select Google Workspace business customers this month. So personal users will have to wait a while for a broader rollout.
Meanwhile, for those who want to build their own translation-powered app, the Gemini live API is worth exploring now.
The Gemini 3.5 Live Translation signals Google’s push to make raw language barriers less noticeable by reducing delays between speech and translation and making conversations feel closer to talking with someone who speaks your own language
Source: Google
