Google DeepMind ships Gemini 3.8 Live with extended thinking variant

Google DeepMind has released Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking, two voice-capable models supporting real-time spoken interactions across 97 languages. The Live model adds audio, visual grounding, and background tool calls while conversations continue, while the Extended Thinking variant enables multi-step reasoning running in parallel with speech generation.

Sources

X mentions

27k ▲

First seen

16Dago

Velocity

+8%/6h

CONTRIBUTING SOURCES

5 ARTICLES

X DISCOURSE

27k TOTAL · TOP 3

@GoogleDeepMind1H · 412.1K

Introducing Gemini 4 Argon – our new frontier model. It’s built for complex workflows across coding, enterprise knowledge work, and cybersecurity defense – rolling out today to a set of trusted testers through our Fairwind Program.

@GoogleDeepMind17H · 438.2K

@demishassabis2D · 427.9K

RT @sundarpichai: Can our TPUs survive and operate in space? Well, we're going to find out. Project Suncatcher is hitching a ride aboard @…