Google Live Translate Supports Over 70 Languages

As a frequent traveler, I have many screenshots of semi-useful translations on my phone. So I was excited when Google announced Gemini 3.5 Live Translate, a new audio model that offers near real-time, voice-to-voice translation in more than 70 languages. Best of all, the model doesn’t wait patiently until the other person finishes speaking; it translates on the fly. While there’s a slight delay (a couple of seconds), the model keeps up in even noisy environments.

What makes this announcement exciting for me is not the number of supported languages. Instead, it’s the sound quality of the translations. The model retains the speaker’s tone, speed, and pitch. This means that the translation has the same rhythm as the original — no more monotonous robot voice. According to Google, “the key to continuous speech synthesis is balancing the trade-off between waiting for context to improve quality and translating immediately to stay in sync.” And that balance between accuracy and delay is crucial in live translation. Google is putting its faith in staying in sync.

1Where you'll actually get it

This isn’t a lab experiment. Google is releasing Gemini 3.5 Live Translate to the public as part of the Google Translate app for Android and iOS worldwide. New for Android is “listening mode,” which plays the translation through your device’s earpiece. This lets you listen to conversations in real time as they’re being translated. Meanwhile, the update to Google Meet boosts support from just five languages with English-only output to more than 2,000 language pairs. That moves live translation from a gimmick to a usable tool for international teams. The Gemini Live API and Google AI Studio provide access to the technology for developers today. Enterprise users will be able to try it out in a private preview of Google Meet later this month.

If you’re not a developer or a diplomat, why should you care? Language barriers are among the final frontiers of daily life. Whether it’s discussing medical issues with a foreign doctor, dealing with a landlord, meeting your significant other’s family, or negotiating with a vendor in another country, effective communication is essential. While written translation has been satisfactory for some time, maintaining a human rhythm during live conversations is a game changer. Rather than awkward, draining interactions, conversations become much more fluid.

A final note: All audio outputs include SynthID, Google’s invisible watermark for AI-generated media. As synthetic voices become a vehicle for fraud, adding authentication to the audio stream is a safeguard. It’s also good to know that this feature is available.

While I don’t think this will replace learning a language (you still need context, humor, and nuance), it does lower the barrier for millions of casual interactions where you need to be understood in the moment.