Gemini 3.8 Live: Google's Real-Time Voice Model, Pricing and Limits

What Gemini 3.8 Live is

Gemini 3.8 Live is a Google audio model for real-time conversation, announced in Google's September 2026 updates. Google's model documentation describes it as "the default option for most low-latency voice agent experiences and real-time dialogue", and the model index lists it as stable with the model code gemini-3.8-live. Google announced it together with Gemini 3.8 Live Extended Thinking, a variant for voice interactions that need more background reasoning.

Inputs, outputs and limits

According to the model page, Gemini 3.8 Live accepts text, images, audio and video, and returns text and audio. The input token limit is 131,072 and the output limit is 65,536 tokens. Google lists audio generation, function calling, the Live API, search grounding and thinking, with interleaved reasoning, as supported. Caching, code execution, file search, grounding with Google Maps, image generation, structured outputs and URL context are listed as not supported. The page gives September 2026 as the latest update and does not state a knowledge cutoff.

Pricing

Google's pricing page lists one price table for Gemini 3.8 Live, Gemini 3.8 Live Extended Thinking and Gemini 3.1 Flash Live Preview. On the paid tier, text input is $0.75 per million tokens and text output $4.50, with output prices including thinking tokens. Audio input is $3.00 per million tokens or $0.005 per minute, audio output $12.00 per million tokens or $0.018 per minute, and image or video input $1.00 per million tokens or $0.002 per minute. The free tier lists input and output as free of charge. Grounding with Google Search includes 5,000 free requests a month shared across Gemini 3 and newer models, then $14 per 1,000 requests.

Where it is available

Google states Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking have rolled out for developers in Google AI Studio and the Gemini API, and for everyone across Search Live, Gemini Live and Google Workspace. Google's update post also describes new expressive voice models introduced with Gemini 3.8 Live.

Status in Google's model index

Google's Gemini API model index lists Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking as new stable models. The same price table also covers Gemini 3.1 Flash Live Preview, an earlier Live model that remains listed as a preview. Output prices on the paid tier include thinking tokens, so the interleaved reasoning Google lists as supported on Gemini 3.8 Live is billed as text output. Search grounding is the only grounding option listed as supported; grounding with Google Maps is not.

How it differs from the Extended Thinking variant

Google's model index describes Gemini 3.8 Live as the default Live API model for low-latency voice agents "without reasoning delays", and the Extended Thinking variant as the higher-reasoning model for when more background reasoning is required. Both share the same input and output types, the same 131,072-token input limit and 65,536-token output limit, and the same price table. On the Extended Thinking model, Google lists function calling as asynchronous only and the Batch API as not supported.

Official Resources