Google Assistant Becomes Smarter and More Agile Thanks to the Gemini 3.8 Live and 3.8 Live Extended Thinking Models

Real-time voice interaction with AI assistants is one of the areas where Google is focusing its efforts to deliver a smoother user experience. Its latest breakthrough in this space is the newly announced Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking models.

Like their predecessors, these voice assistants enable fluid, natural, real-time voice conversations, but they stand out for delivering significant advances in their reasoning capabilities and background task execution, as well as greater conversational fluency.

As a result, having spoken conversations with Gemini Live through the Gemini app, Google Workspace, or Search becomes a faster, more intuitive, and more collaborative process, in which the AI can carry out and resolve more complex tasks more efficiently.

Introducing our most advanced Gemini Audio models yet 🗣

Gemini 3.8 Live and 3.8 Live Extended Thinking let you speak, collaborate, and execute tasks seamlessly, meaning conversing with AI just got a lot more natural.

So, what’s the difference between these two models? Let’s… pic.twitter.com/Dfy7j4zxOh

— Google AI (@GoogleAI) September 15, 2026

Gemini 3.8 Live

Gemini 3.8 Live is optimized for everyday, fluid interaction, as it has been designed to be fast, cost-effective, and efficient. The model has a strong visual foundation that allows it to process visual inputs virtually in real time, adding contextual information to its system and helping it deliver more accurate responses and results.

It offers smooth conversation, with the ability to handle interruptions mid-sentence, and it can detect up to 97 languages, switching from one supported language to another within the same conversation if the user needs it. In addition, it is multitasking: while conversing, it can run tools and API calls in the background to fulfill the tasks and requests assigned to it.

Who can use Gemini 3.8 Live

Google has already begun rolling out Gemini 3.8 Live in the Gemini API and in Google AI Studio for developers, in the private preview of Gemini Enterprise for businesses, and in Search Live for the general public.

The tech company also announced that version 3.8 Live will soon arrive in the Gemini Enterprise customer experience as well, although it has not provided a specific date.

Gemini 3.8 Live Extended Thinking

For its part, Gemini 3.8 Live Extended Thinking is designed for use in highly complex tasks that require greater intelligence and deeper multi-step reasoning. This model can maintain fluid, uninterrupted conversations while executing complex workflows. It also processes and understands visual information within seconds.

While it works, it produces “early verbal cues such as ‘Let me check that…’ to naturally acknowledge instructions, along with real-time narration that guides users through multi-step background tasks as they progress,” Google explains.

Who can use Gemini 3.8 Live Extended Thinking

Gemini 3.8 Live Extended Thinking is already available to developers in the Gemini API and in Google AI Studio. Businesses, meanwhile, can access the model through a private preview in Gemini Enterprise (it will also soon be activated in Gemini Enterprise for Customer Experience and Google Workspace enterprise customers).

In addition, it is already available to the general public in Gemini Live, to Google AI Pro and Ultra subscribers in Workspace in Docs, and to all Google AI subscribers in Gmail and Keep.

SynthID watermark

All audio generated by these Gemini Live models includes SynthID, a digital watermark imperceptible to the human ear embedded directly into the audio output, making it possible to verify and identify when a voice file has been generated by Google AI.

SynthID is not exclusive to Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking, but is also integrated into other Google AI products such as Nano Banana and VEO3.

Photo: Google