Overview

• Google Meet's real-time speech translation feature launched in beta on May 20, 2025, for AI Pro and Ultra subscribers.
• The tool currently supports English and Spanish, with Italian, German, and Portuguese to follow in the coming weeks.
• DeepMind's large language audio model powers the feature, preserving voice tone, expression, and cadence for natural conversations.
• Translated speech is overlaid on the speaker's faint original voice, enabling seamless multilingual communication with low latency.
• Enterprise users will gain access to early testing later this year as Google continues its phased rollout.