Gemini 2.5 Flash-Lite is a streamlined reasoning model from the Gemini 2.5 family, designed for extremely low latency and cost-effectiveness. It delivers higher throughput, quicker token generation, and enhanced performance on standard benchmarks compared to previous Flash models.