Google Launches Gemini 3.5 Live Translate-Real-Time Voice Translation Across 70+ Languages

Language barriers in live conversation may soon be a thing of the past. Google has unveiled its most advanced translation model yet. It does not just translate words — it preserves your voice.

What Is Gemini 3.5 Live Translate?

Google announced the launch of its new Gemini 3.5 Live Translate model for real-time speech-to-speech translation. The goal is to reduce language barriers and enable more natural, seamless live conversations between users speaking different languages. 

This is not a simple upgrade. It is a fundamental rethink of how translation technology works.

It Translates While You Speak — Not After

Old translation tools made you wait. You spoke. The tool processed. Then it translated.

Gemini 3.5 Live Translate works differently. The model continuously generates translations during speech, instead of waiting for the speaker to finish a sentence before starting the translation process.  

The result feels closer to a real conversation. No awkward pauses. No broken rhythm.

Your Voice Stays Your Voice

This is one of the most impressive parts of the new model. Gemini 3.5 Live Translate automatically detects more than 70 languages while preserving the speaker's intonation, pacing, and pitch.  

Your tone of confidence, your warmth, your emphasis — none of it gets lost in translation. The model carries your personality across the language divide.

Real-World Uses Google Demonstrated

Google did not just talk about the technology. They showed it.

Google showcased the model through practical examples including real-time video dubbing, as well as translating multilingual lectures and presentations — all while maintaining a smooth conversational flow between participants.  

Imagine a business meeting where every person speaks their own language. Or a university lecture reaching students in 20 countries simultaneously. That is the direction this technology is heading.

Google Meet Gets a Major Upgrade

The impact will soon reach millions of everyday users.

Google confirmed that Gemini 3.5 Live Translate will be integrated into Google Meet later this year. The number of languages supported for live translation will increase from five to over 70, with the potential to enable more than 2,000 different language combinations in a single meeting. 

The service will also move beyond its previous reliance on English as an intermediary language for translation. Google also plans to update the Google Meet interface to make it easier to access the voice translation feature more quickly. 

The improved version will first become available to a select group of companies participating in Google Workspace as part of a special preview later this month. It will then expand to a wider range of users later this year. 

Already Rolling Out on Android and iOS

You do not have to wait for Google Meet. The rollout has already started.

Google has begun rolling out Gemini 3.5 Live Translate globally within the Google Translate app on both Android and iOS. The update includes a new listening mode on Android, which allows users to hear live subtitles through headphones or directly through the phone's speaker. 

Built-In Safety Against Misuse

Google is not ignoring the risks of AI-generated audio.

All audio clips generated by the model will carry an invisible digital watermark using SynthID — a technology developed by Google to help verify AI-generated content and limit its misuse in spreading misinformation. 

Every translated voice output will be traceable. That is a meaningful safeguard in an era of deepfakes and synthetic media.

At a Glance

Real-time voice translation that sounds like you — in any language. That is what Google is building. And it is arriving faster than most people expected.