Language barriers are one of the biggest challenges in a globally networked world. With Gemini 3.5 Live Translate, Google is now presenting a new AI-supported solution for real-time language translation based on Gemini 3.5. The model translates spoken language directly during a conversation, automatically recognizes more than 70 languages and even preserves tone of voice, pace of speech and voice characteristics. The technology is designed to make communication across language barriers more natural and fluent – for private individuals as well as for companies and developers.
- Real-time voice translation with support for over 70 languages.
- Natural reproduction of tone of voice, pace of speech and pitch.
- Use in Google Translate, Google Meet and via the Gemini Live API.
- New security feature with SynthID watermarking for AI-generated audio data.
Gemini 3.5 Live Translate enables more natural conversations in real time
Machine translation has made enormous progress in recent years, but many solutions still feel artificial or delayed. This is exactly where Gemini 3.5 Live Translate comes in. Instead of relying on traditional “speak-wait-translate” methods, the new audio model processes speech continuously during the conversation.
This results in much smoother conversations with minimal delay. According to Google, the translation lags just a few seconds behind the speaker without noticeably interrupting the flow of the conversation. At the same time, the system automatically recognizes the language used and does not require any manual configuration.
Another focus is on the naturalness of the rendering. While many translation services correctly translate the content, emotional nuances are often lost. Gemini 3.5 Live Translate, on the other hand, attempts to preserve intonation, speech rhythm and voice pitch as authentically as possible. This makes conversations seem more natural and personal.
You are currently viewing a placeholder content from YouTube. To access the actual content, click the button below. Please note that doing so will share data with third-party providers.
Technical robustness has also been improved. The model can work in noisy environments and is therefore suitable for meetings, events, teaching situations or conversations on the move. This development could significantly simplify communication, especially in international teams or global business relationships.
Gemini 3.5 Live Translate comes to Google Meet and Google Translate
Google is gradually integrating the new technology into several products. Developers already have access to the functions via the Gemini Live API and Google AI Studio. This makes it easier to implement applications for live interpreting, multilingual conferences or international communication platforms.
There are also new opportunities for companies. In Google Meet, Gemini 3.5 Live Translate will enable language translations in more than 70 languages in future. Whereas previously only a few language combinations were supported, in future over 2,000 language combinations will be possible within a video conference. Google is thus significantly expanding the areas of application for international meetings.
At the same time, the global rollout is starting in the Google Translate app for Android and iOS. Users can use the live translation function with headphones and receive the translated language almost in real time. Google is also introducing a new listening mode for Android devices. The translation is played back directly through the phone speaker at the ear – similar to a traditional phone call. This can be particularly helpful for city tours, travel or spontaneous conversations.
Google also adds SynthID to all audio files generated by the AI. This invisible digital watermark is intended to help make AI-generated content recognizable and prevent misuse or disinformation.
Conclusion
With Gemini 3.5 Live Translate, Google is taking a significant step towards natural, real-time communication between different languages. The combination of low latency, automatic speech recognition and the preservation of natural language features sets the solution apart from many previous translation systems. The technology is now being rolled out gradually via Google Translate, Google Meet and the Gemini Live API. Information on additional costs has not yet been published.
Source: Google
