Models/press release/2026-09-15

Google DeepMind announces Gemini 3.8 Live with simultaneous thinking and speaking

Google DeepMind has introduced Gemini 3.8 Live, a new language AI model. According to the company, this model is designed to think and speak simultaneously, incorporating "Extended Thinking" capabilities. The primary claims regarding its simultaneous processing and advanced reasoning come directly from Google DeepMind's own announcement.

What every outlet reports

  • Google DeepMind introduced Gemini 3.8 Live
  • The model is capable of simultaneous thinking and speaking
  • It features "Extended Thinking" capabilities

All 2 articles

optimist1
Introducing Gemini 3.8 Live and 3.8 Live Extended ThinkingoriginalGoogle DeepMind Blog · company statement

Related stories