FrontierThe story, in brief
Google’s new speech model Gemini 3.8 Live supports real-time reasoning
Google ships Gemini 3.8 Live with real-time reasoning — solving the latency wall that's kept voice agents from shipping.
The KeyNews take
Why it matters
A major frontier lab addresses a critical technical bottleneck in agentic voice AI. Real-time reasoning in speech models is a capability milestone that directly enables the next wave of voice agent products.
The key facts
6 to knowGemini 3.8 Live and Gemini 3.8 Live Extended Thinking released
Marketed as Google's most advanced voice processing models to date
Near-real-time reasoning capability
Simultaneous speech-and-thought processing
Latency reduction for voice-based AI agents is the stated problem being solved
Published September 15, 2026
Go to the source
SiliconAnglesiliconangle.com
Publisher excerpt: Google LLC is trying to address the latency problem associated with voice-based artificial intelligence agents with the launch of Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking today. They’re billed as the company’s most advanced voice processing models released so far, and they’re capable…