gemini-3.5-flash-lite

gemini-3.5-flash-lite is a lightweight model in the Gemini 3.5 Flash series, targeting scenarios where a large number of repetitive requests need to be completed at a low resource cost. It is suitable for content preliminary screening, short summaries, field standardization, language identification, text routing, template reply and candidate result sorting. Placing it at the front end of a multi-model process can filter simple requests before handing off difficult or high-value tasks to more capable models. Before going online, real traffic samples should be used to test the misclassification rate, output consistency and abnormal input performance, and a manual sampling inspection mechanism should be retained.

Context

1048576 tokens

Max output

65536 tokens

Input

llm image video text-to-speech

Output

llm