Google releases Gemini 3.8 Live and 3.5 Transcribe audio models for real-time voice apps

gaganghotra_ · x · 2026-09-16

Google AI Studio announced new Gemini audio models — Gemini 3.8 Live and 3.5 Transcribe — now available via the Gemini API and Google AI Studio, enabling developers to build real-time conversational voice applications.

Related event: Google Launches Gemini 3.8 Live and Extended Thinking, Topping Speech Quality Index(21 posts)→

Original post →

More from Multimodal

Multimodal channel →