Home / Companies / Box / Blog / Post Details
Content Deep Dive

Gemini 3.5 Flash-Lite on real enterprise work

Blog post from Box

Post Details
Company
Box
Date Published
Author
Sidharth Srinivasan, Product Manager, Box AI
Word Count
608
Company Posts That Month
38
Language
English
Hacker News Points
-
Post removed?
No
Summary

Box evaluated Gemini 3.5 Flash Lite using the Box Complex Work Eval, which benchmarks document-grounded tasks across various industries, and found it significantly outperformed its predecessor, Gemini 3.1 Flash Lite. Gemini 3.5 Flash Lite scored 58% overall compared to the previous model's 41%, demonstrating a 17-point advantage, particularly in complex tasks involving multi-step quantitative reasoning and imperfect data. This model excelled in industries that rely heavily on numerical analysis, such as Consumer Products, Retail, Financial Services, and Legal, delivering more accurate and defensible results. Additionally, Gemini 3.5 Flash Lite maintained its accuracy across different types of analytical work, showing the greatest improvement in data analysis tasks and handling chained calculations effectively. Notably, it achieved these results with greater efficiency, taking roughly a quarter of the time and using fewer resources than the previous version, making it a more accurate and resource-efficient option for enterprise workloads. Gemini 3.5 Flash Lite will soon be available through Box AI, offering enhanced performance for complex enterprise knowledge work.

Trends Found in this Post

No tracked trend matches for this post yet.

Use This Data

Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.