LLM · Google

Gemini 3.1 Flash Live Preview

Google's model page lists gemini-3.1-flash-live-preview as a legacy preview with Live API supported, inputs of text, images, and audio, and output of text and audio. The input token limit is 131,072 and the output token limit is 65,536. Caching and batch are not supported. The page recommends Gemini 3.8 Live. The shared pricing table lists the same text rates, $0.75 input and $4.50 output per 1M tokens, plus audio input at $3.00, audio output at $12.00, and image or video input at $1.00. The stored figures are the text rates.

Caching and batch

Live API Yes
Cache read Not supported
Cache write Not supported
Cache TTL Not supported
Minimum cacheable tokens Not supported
Batch discount Not supported

The pricing table groups gemini-3.1-flash-live-preview with gemini-3.8-live. Text input is $0.75 and text output is $4.50 per 1M tokens. Audio input is $3.00 and audio output is $12.00. Image or video input is $1.00. The model page lists caching and batch as not supported, and calls this a legacy preview.

Checked 25 Oct 2026 Source

Checked 25 Oct 2026https://ai.google.dev/gemini-api/docs/models/gemini-3.1-flash-live-preview