Google Ships Gemini 3.5 Live Translate: Real-Time Voice Translation in 70+ Languages That Keeps Your Voice
Google's new speech-to-speech model translates conversations continuously across 70+ languages while preserving the speaker's tone, tempo, and pitch. It's rolling out globally on Android and iOS, with every output watermarked by SynthID.
Google has released Gemini 3.5 Live Translate, a dedicated speech-to-speech model that translates spoken conversation in real time across more than 70 languages — and it’s rolling out globally on Android and iOS right now.
The technical shift matters more than the language count. Older translation systems wait for a speaker to finish a sentence, then read back a flat, synthesized rendering. Gemini 3.5 Live Translate processes speech as it flows, staying just a few seconds behind the speaker. And it preserves intonation, tempo, and vocal pitch, so the translated voice sounds like you speaking another language rather than a robot summarizing what you said.
Android gets an exclusive trick: Listening Mode. Hold the phone to your ear like a regular call and hear a private translation through the earpiece — no earbuds, no awkward speaker-phone huddle. For travelers and anyone working across language barriers, that’s the kind of friction removal that makes a feature actually get used.
Watermarked by default
Every audio output carries a SynthID watermark — an inaudible marker that flags the audio as AI-generated. The timing is deliberate. Article 50 of the EU AI Act, which requires labeling of synthetic content, comes into force on August 2, 2026. Google is shipping compliance two months early and turning a regulatory obligation into a trust feature. Expect competitors to follow under less flattering circumstances.
Developers get it too
The model is available now in public preview through the Gemini Live API and Google AI Studio. Grab, Agora, and LiveKit are already building on it — think ride-hailing drivers talking to tourists, or video calls that translate themselves. Google Meet users on Workspace plans get a private preview starting this month, with a broader rollout later in the year.
The bigger picture
This is the second Gemini 3.5 release in a month, following Flash’s general availability at I/O in May, and it lands while the flagship Gemini 3.5 Pro — with its 2-million-token context and Deep Think reasoning — is still pending. Google is shipping the family piece by piece, and the pieces are landing in products people already use daily: Translate, Meet, Android itself.
Real-time translation has been demoed at keynotes for a decade. The difference now is that it’s shipping globally, it keeps the speaker’s voice, and developers can build on it today. That combination didn’t exist last week.