DeepMind announces Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking, describing them as its most advanced “live” dialogue models. The company positions the releases as improvements for natural, back-and-forth conversation in real time.
The announcement emphasizes that both models are designed to support conversational interaction, with “Extended Thinking” suggesting an additional approach intended to deepen how the system handles dialogue. In the coverage provided, The Keyword and DeepMind both describe the same core products and their focus on live dialogue quality, without citing detailed benchmark results or specific performance claims.
Across the sources, the shared angle is the introduction of two new model variants and the intention to improve how Gemini responds in ongoing conversation. Neither source excerpt includes technical specifications, availability details, pricing, or quantified evaluations, so differences in framing are limited to the shared emphasis on natural conversation and live operation.
