Gemini 3.5 Live Translate [Courtesy of Google]Google has unveiled a translation model capable of real-time interpretation in over 70 languages.

Google announced on the 9th (local time) that it is fully implementing its new translation model, "Gemini 3.5 Live Translate," based on its latest artificial intelligence (AI) model, into the Google Translate app.

Google introduced that the new translation model has shifted from sequential interpretation to a "continuous real-time generation" method, which is close to simultaneous interpretation.

This means the model balances immediate translation of speech with waiting to improve quality and understand context.

As a result, the translated audio has a time lag of only a few seconds compared to the original speaker's words.

Furthermore, whereas users previously had to pre-select the languages to be translated, the new model automatically detects and translates languages. It can also be used in multilingual conversation environments where multiple languages are mixed.

The speech generation technology itself has also been enhanced, preserving the original speaker's intonation, speaking style, and pitch to deliver natural-sounding audio.

Previously, using voice translation in the Google Translate app on iPhones and Android phones required connecting earphones. However, now, even without earphones, users can hear translated audio by holding the smartphone to their ear, similar to a phone call.

Google plans to apply this feature to its video conferencing platform, "Google Meet," to support multilingual meetings.

Major service companies such as Grab and Agora, which have pre-tested the model, have praised its capabilities.

Baek Hyun-jung, Chief AI Officer (CAIO) at CJ ENM, who participated in the initial testing, also commented, "It demonstrates outstanding quality, capable of providing a more vivid experience to global and Korean viewers."