Google has launched Gemini 3.5 Live Translate, an AI model that provides real-time, continuous speech translation in over 70 languages, preserving vocal characteristics. The technology is being integrated into Google Meet and Google Translate, significantly expanding their multilingual communication capabilities. Google is also offering the technology via the Gemini Live API and Google AI Studio for developers, with partners already integrating it for various real-time translation and dubbing applications.
The shift from turn-based to streaming translation drastically reduces conversational latency, removing the 'stop-and-start' friction typical of legacy live interpretation. For the streaming ecosystem, this facilitates low-cost, real-time dubbing for live broadcasts, webinars, and education—a market segment projected to reach $1.35 billion in 2026. By making this accessible via API at a competitive price point, Google is challenging incumbents like Meta and OpenAI in the race to provide the underlying infrastructure for globalized live video. Industry observers should track adoption rates among mid-tier streaming platforms and live event producers through late 2026.
Google's Gemini 3.5 Live Translate arrives as the company has steadily expanded the Gemini Live API from an experimental real-time interaction tool into a production-ready platform for streaming audio applications. The Live API entered preview in April 2025 with support for 30 new languages, configurable voice activity detection, and virtually infinite sessions through a sliding context window, laying the infrastructure groundwork that now hosts the dedicated translation model. The progression from general-purpose streaming audio to a purpose-built translation model signals Google's intent to capture developer mindshare in the live localization space before competitors establish entrenched API ecosystems.
Google has positioned the translation capability across multiple product surfaces simultaneously. The Gemini 2.5 Native Audio update introduced live speech-to-speech translation preserving intonation, pacing, and pitch, rolling out first in the Google Translate app as a beta for headphone-based continuous listening and two-way conversation. That release established the consumer-facing foundation, while Gemini 3.5 Live Translate extends the same architecture to enterprise and developer use cases through Google Meet private preview and the public Gemini Live API. The model handles multilingual inputs without manual language configuration and maintains noise robustness for unpredictable live environments, both critical requirements for broadcast and event production workflows.
For developers building real-time translation into streaming applications, the Gemini Live API provides specific configuration controls that affect production behavior. The API's translation settings include a targetLanguageCode parameter using BCP-47 language codes and an echoTargetLanguage boolean that determines whether input already in the target language is repeated or silenced, giving integrators granular control over output in multilingual sessions. The model generates speech continuously rather than waiting for a speaker to finish, balancing context accumulation against synchronization delay, which Google describes as keeping output just a few seconds behind the speaker throughout a session.
Add StreamingMeme as a preferred source on Google to see more of our streaming news at the top of your Search results.
Add as preferred source