Google completes Gemini 3.1 tier with Flash-Lite: A high-speed, low-cost utility for enterprise scale

Google launches Gemini 3.1 Flash-Lite, offering 2.5X faster token delivery and 1/8th the cost of Gemini 3.1 Pro for enterprise-scale AI.

By: AXL Media

Published: Mar 5, 2026, 5:49 AM EST

Source: The information in this article was sourced from VentureBeat

Technical Performance and Benchmarks

Gemini 3.1 Flash-Lite is optimized for high-throughput enterprise tasks where latency is the primary barrier to user experience. It achieves an output speed of 363 tokens per second, significantly outpacing Gemini 2.5 Flash.

Despite its "Lite" designation, the model maintains high cognitive scores across several standardized benchmarks:

MMMLU (Multilingual Q&A): 88.9%

GPQA Diamond (Scientific Knowledge): 86.9%

MMMU-Pro (Multimodal Understanding): 76.8%