Voice agent builders get 97 languages and background tool calls at the same price as the outgoing 3.1 Live, with a free tier.

Kenning's take

What it does

Gemini 3.8 Live is a pair of real-time voice models. The standard version is built for cost and scale, while 3.8 Live Extended Thinking reasons through harder tasks while it keeps talking and narrates its progress. They switch between 97 languages mid-conversation and run tool calls in the background.

• Two versions, standard and Extended Thinking
• Switches between 97 languages
• Background tool calls during conversation
• Near real-time visual input
• SynthID watermark on audio

Who it suits. Developers building voice agents and people using Gemini Live or Search Live.

Key facts

Category

AI models

Part of

Google Gemini

Made by

Google

Based in

United States

Released

15 September 2026

How to get it

Gemini API and AI Studio; Search Live; Extended Thinking in Gemini Live and some Workspace apps for Google AI subscribers; Gemini Enterprise private preview

Works on

Web, Android, iOS, API

In the UK

Yes

Free plan

Yes, Gemini API free tier

API price

$0.75 text / $3.00 audio input, $4.50 text / $12.00 audio output per million tokens

FAQ

Facts from the maker's own pages and the sources listed. Checked 2026-10-11.