Google has released Gemini 3.5 Live Translate, an audio translation model designed to translate speech continuously rather than waiting for each sentence to finish. The system supports more than 70 languages and aims to preserve tone, pace, and pitch.
The immediate product impact is in Google Meet, where language support expands sharply. More broadly, it shows how multimodal AI is moving from text chat into live communication workflows.
If latency and voice preservation hold up in real meetings, translation could become one of the clearest everyday uses for frontier audio models.