Following the main artificial intelligence model Gemini 3.8 Flash, Google release Its additional versions are Gemini 3.8 Live and Live Extended Thinking. They are designed as ready-to-use voice assistants that make communication with Gemini more comfortable and collaborative.
Image source: blog.google
Google explained that the Gemini 3.8 Live model is designed to work in scalable and efficient scenarios, while the Gemini 3.8 Live extended thinking is designed to solve high-complexity problems. The younger model can process visual data almost instantly; it can handle switching between 97 supported languages during a conversation. It works with external tools and handles API calls in the background, acknowledging requests without interrupting the conversation with the user.
One of the distinguishing features of Gemini 3.8 Live Extended Thinking is the ability to reason and speak simultaneously. Google emphasized that it proved powerful enough to rank first in human-analyzed Voice Quality Index model ratings; and also became a leader in the τ-Voice-banking benchmark, which involves testing voice AI agents while performing complex tasks in real-time in the banking technical support and customer support areas of fintech companies.
The sounds produced by the model will receive the SynthID digital watermark. Ordinary users will be able to use Gemini 3.8 Live models in the Search Live service; Extended Thinking in Gemini 3.8 Live – In Gemini Live. AI Pro and Ultra paying subscribers will be able to access Gemini 3.8 Live Extended Thinking in Google Docs, Gmail and Keep. For developers, both models are available through the Gemini API and Google AI Studio platform. For enterprise users, these two modes can be accessed through Gemini Enterprise; in the future, the new mode will appear in Gemini Enterprise for customer experience services and Google Workspace for enterprise customers.
If you find an error, select it with your mouse and press CTRL+ENTER.
