Google has released Gemini 3.5 Live Translate, a new AI model that provides real-time voice-to-voice translation across more than 70 languages. The update expands on previous efforts to bring live translation to more users, with the model designed to keep up with normal conversation pace and match intonation, pacing, and pitch. According to Google, the new model is faster than its predecessor, with lower latency, allowing for near real-time translation. The feature is rolling out across several parts of the Google ecosystem, including Google Meet and the Google Translate app on both Android and iOS.
The Gemini 3.5 Live Translate model is part of the version 3.5 family launched at I/O, with a Flash version already available and a Pro model expected soon. It processes speech continuously and handles multilingual inputs automatically, eliminating the need for manual configuration. The model also filters out background noise in busy environments. Developers can access a public preview through the Gemini Live API or AI Studio. Select enterprise customers will get access to the new translation model in Google Meet starting this month, with the interface being tweaked to bring the live translate feature to the front.
Google began testing Gemini-based live translation in the app with any earbuds last year, previously requiring Pixel Buds with an Android phone. The pending update will expand further with the addition of the latest 3.5 model. Users can now use any earbuds or, if none are available, hold the phone up to their ear like they’re on a call to hear a near real-time English translation of a guided tour in Spanish. The audio streams from Gemini 3.5 Live Translate are intended to sound lifelike but are marked with SynthID watermarks, which cannot be removed.
Source: arstechnica
