Gemini 3.1 Flash Live, our highest-quality audio and voice model, is launching today! This is how it advances our real-time dialogue capabilities:
— Faster: 3.1 Flash Live powers faster responses than the previous model, which is great for when you need a timely answer (Ex: “Can you help me change this tire in under 5 minutes?!?”)
— Longer: In Gemini Live, the model’s context window is now twice as long as before, so it can keep up all of the details shared in your conversations (Ex: “I'm back to writing my future bestselling crime novel. Remind me, who is the secret double agent?”)
— Global: 200+ more regions will be able to have real-time, multimodal conversations in their preferred language
Beginning today, you can try out 3.1 Flash Live in these places:
🔊: Gemini Live in @GeminiApp + Search Live
🛠️: Gemini Live API + @GoogleAIStudio in preview
💼: Gemini Enterprise for Customer Experience
context window — The maximum amount of text a model can consider at once — its working memory for the current conversation or task.
Why it matters
Voice agents on the Live API get longer conversational memory and faster responses, and the expanded region list removes a common blocker to shipping realtime voice outside a few markets.
