Google Unveils Gemini 3.5 Live Translate for Real-Time Conversations in 70+ Languages

Google has introduced Gemini 3.5 Live Translate, a new AI-powered speech translation model designed to make multilingual conversations smoother and more natural. The company says the feature supports real-time speech-to-speech translation in more than 70 languages, aiming to eliminate delays that are common in existing voice translation systems.

Traditional translation tools typically follow a turn-by-turn process, where users must wait for a speaker to finish before the translated response is generated. According to Google, Gemini 3.5 Live Translate takes a different approach by processing speech instantly as users speak, allowing translated audio to flow continuously during conversations.

Google explained that the model is designed to maintain conversational rhythm without introducing long pauses or interruptions. The company added, “No awkward pauses or choppy audio, just real connection without language barriers.”

The new system has also been developed to preserve elements such as intonation, pacing, and speaking style. This is expected to create a more natural listening experience compared to the robotic tone often associated with older translation tools.

While Apple already offers a Live Translation feature, its language support currently remains limited. Google, on the other hand, claims Gemini 3.5 Live Translate can handle over 70 languages, significantly expanding multilingual communication possibilities.

Google also confirmed that all audio produced through its AI models includes SynthID watermarking technology. The company says the watermark is embedded directly into the generated audio and remains imperceptible to listeners while still being detectable for verification purposes.

Alongside Live Translate, Google has introduced a new Listening Mode for the Google Translate app. The feature is designed for situations where users may not have access to headphones. Once activated, users can hold the phone near their ear like a standard phone call and listen to translated speech through the device’s earpiece.

The company is also extending the technology to Google Meet. Previously, Google Meet supported speech translation in only five languages. With Gemini 3.5 Live Translate, support is expanding to over 70 languages with real-time translation capabilities.

Google stated that the update will enable more than 2,000 language combinations within a single meeting, removing the earlier limitation of translating only to and from English. The Meet interface is also being updated to provide quicker access to live speech translation features during meetings.

In terms of availability, Google says Gemini 3.5 Live Translate has already started rolling out on the Google Translate app for both Android and iOS users. Developers can access the model through the Gemini Live API and Google AI Studio in public preview. Meanwhile, Google Meet integration will initially arrive in private preview for select Workspace users starting this month.

Google noted that its translation services now process more than a trillion words every month across its products for users worldwide.