Gemini 3.8 Live
Gemini 3.8 Live handles live text, image and audio input with text and audio output.
What the release provides
Interleaved reasoning is built in; requests should omit thinking_level and thinking_config. The API supports synchronous and asynchronous function calls and proactive audio. Batch requests and structured output are not supported.Gemini 3.8 Live official API documentation
How to use it well
Practical guidance. Design the conversation around interruptions, pauses and delayed tool results. Keep the audio stream responsive while an action runs. Validate tool arguments on the application side and make the distinction between a spoken intention and a completed action clear.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Access | API |
| Model identifiers | gemini-3.8-live |
| Context window | 131,072 tokens API input limit; reserve output capacity where the service shares one window. Gemini 3.8 Live official API documentation |
| Maximum output | 65,536 tokens Published API output limit; request settings can be lower. Gemini 3.8 Live official API documentation |