Google Research· ResearchNew model

Real-time speech-to-speech translation

Google Research introduces an end-to-end speech-to-speech translation model that enables real-time translation in the original speaker's voice with only a 2-second delay, using a streaming architecture and a scalable data acquisition pipeline.

From the source

We introduce an innovative end-to-end speech-to-speech translation (S2ST) model that enables real-time translation in the original speaker's voice with only a 2-second delay — bringing long-imagined technology into reality and making cross-language communication more natural.

research.google

Announced

19 Nov 2025 09:59 UTC

Detected

03 Aug 2026 21:12 UTC (+257d after)