News/Models
Try nowModels·MeaningfulOfficialbreakingUpdated Sep 15·Updated 2×·First seen Sep 15

Gemini 3.8 Live and Extended Thinking Models Launched

LaunchTrending
Try it

Something you can actually use or run today.

On September 15, 2026, Google DeepMind launched Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking for integration in Google AI Studio.

It provides a direct alternative to OpenAI's real-time and reasoning models, preventing a total market monopoly on inference-time compute scaling.

AILookup take

Google is finally matching the feature set of the 'o1' class of models. The native integration into Google AI Studio makes this a compelling choice for builders already locked into the GCP ecosystem.

Who cares
GCP developersvoice assistant buildersAI researchers
Watch next

A head-to-head coding benchmark comparison between Gemini 3.8 Extended Thinking and GPT-6 Astra.

Details
  • Builders gain access to lower-latency conversational voice endpoints via the updated Gemini Live API.
  • Extended Thinking allows the model to allocate more compute at inference time for complex coding and math.
  • Offers a direct, highly competitive alternative to OpenAI's real-time and reasoning model suites.
Consensus

Broad agreement that this updates Google's production conversational and reasoning capabilities significantly.

Introducing Gemini 3.8 Live and 3.8 Live Extended ThinkingYouTube: Google for DevelopersHacker NewsGoogle DeepMind· 5 stories
Tools in this story

Reviewed profiles on AILookup

AILookup

Research utility for AI tools. Compare reviewed profiles, distinguish listed tools from reviewed coverage, and track tool changes without marketing fluff.

© 2026 AILookup. All rights reserved.