Product Launch
Important
Medium
90% Confidence
Google Launches Gemini 3.1 Flash Live Audio Model with Enhanced Real-time Dialogue and Multilingual Capabilities
Summary
Google released Gemini 3.1 Flash Live audio model, scoring 90.8% on ComplexFuncBench Audio with reduced latency and doubled conversation thread duration. It is available via APIs for developers and enterprises, featuring SynthID watermarking and enhanced multilingual support.
Key Takeaways
Google launched Gemini 3.1 Flash Live audio model on March 26, 2026. It scored 90.8% on ComplexFuncBench Audio and 36.1% on Scale AI's Audio MultiChallenge (with 'thinking' mode). Compared to previous versions, it has faster response times, doubled conversation thread duration, and improved tone understanding and multilingual capabilities.
The model is available via Gemini Live API (Google AI Studio) for developers, Gemini Enterprise for businesses, and publicly through Search Live and Gemini Live in over 200 countries. All generated audio includes SynthID watermarking.
The model is available via Gemini Live API (Google AI Studio) for developers, Gemini Enterprise for businesses, and publicly through Search Live and Gemini Live in over 200 countries. All generated audio includes SynthID watermarking.
Why It Matters
Google strengthens its AI voice strategy, promoting enterprise voice agent deployment and potentially intensifying competition with AWS and Azure in real-time AI applications....