Gemini-3.5-Flash-Lite, Google’s fastest cost-efficient multimodal model in the Gemini 3.5 family. It delivers up to 350 output tokens per second with low latency and high throughput. Natively supports text, image, audio & video. Ideal for high-volume workloads including translation, document parsing, lightweight agent tasks and structured data generation. It achieves major improvements over Gemini 3.1 Flash-Lite and supports adjustable thinking levels for scalable production traffic.