The Gemini 3.8 family is getting a little bigger with two new additions

It’s been about two weeks since Google announced Gemini 3.8 Flash, which arrived only three weeks after it launched Gemini 3.7 Flash. The time for a new successor hasn’t come quite yet, but we are getting introduced to new AI models today. While we wait for the next step up, Google is expanding the Gemini 3.8 family with two new additions.

The latest additions to the Gemini 3.8 line are called Live and Live Extended Thinking. Both models are designed to be production-ready voice assistants that make speaking with Gemini more fluid and collaborative. As Google explains in its blog, Live is built for scaling and cost efficiency, while Live Extended Thinking is meant for high-complexity tasks.

Diving into the details, Google says that Gemini 3.8 Live is able to process visual inputs in near real-time. It’s said to be able to detect transitions between 97 supported languages mid-conversation. The model can also handle tools and API calls in the background, allowing it to acknowledge requests while also keeping up with your conversation.

Meanwhile, Live Extended Thinking is able to reason and speak simultaneously. The tech giant claims that Live Extended Thinking is powerful enough that it captured the top spot on Artificial Analysis’ Speech to Speech Quality Index. It also appears to be the leader in agentic task completion and in τ-Voice-banking, which tests real-time, audio-native conversational AI agents on complex, real-world banking and fintech customer support tasks.

Just like previous models, all audio generated by these models will be watermarked with SynthID. This is Google’s digital watermarking and detection tool that embeds invisible signals into AI-generated content.

Both Live and Live Extended Thinking are rolling out today. For average users, you’ll find Gemini 3.8 Live in Search Live, and Gemini 3.8 Live Extended Thinking will be available in Gemini Live. Those with AI Pro and Ultra subscriptions will also find Live Extended Thinking available in Google Docs, Gmail, and Keep.

As for developers, both models will be available in the Gemini API and Google AI Studio. Meanwhile, enterprise users will have access to both models in private preview in Gemini Enterprise. They will also eventually be coming to Gemini Enterprise for Customer Experience and Google Workspace business customers.