Gemini 3.1 Flash-Lite

About Gemini 3.1 Flash-Lite

Gemini 3.1 Flash-Lite is a highly capable natively multimodal reasoning model developed by Google DeepMind. Released in March 2026, it serves as the cost efficient and fast addition to the Gemini 3 series, specifically optimized for high volume latency sensitive tasks. The model is based on the Gemini 3 Pro architecture and delivers exceptional performance while maintaining significantly lower operational costs compared to larger tier models.

Designed for developers and enterprises requiring responsive real time experiences, Flash-Lite offers a 1 million token context window and generates up to 64K tokens in output. It achieves an impressive output speed of 363 tokens per second while outperforming previous generation models like Gemini 2.5 Flash in both speed and quality benchmarks. The model comes standard with adjustable thinking levels, allowing developers to control reasoning depth based on task complexity.

Key Features

• Adaptive Thinking: Adjustable reasoning levels let developers control how much the model thinks per task, balancing speed and depth for high frequency workflows.
• Multimodal Processing: Native support for text images audio and video inputs with sophisticated understanding across modalities.
• High Speed Output: Delivers 363 tokens per second with 2.5X faster time to first token than previous Flash versions.
• Cost Efficiency: Priced at $0.25 per million input tokens and $1.50 per million output tokens for economical high volume operations.
• Large Context Window: Supports up to 1 million input tokens and 64K output tokens for extensive document and video analysis.
• Strong Benchmarks: Achieves 86.9% on GPQA Diamond and 76.8% on MMMU Pro, surpassing larger models from prior generations.