Google's Gemini 3.7 Flash Model Brings Improved Performance to Cloud-Based Agent Spark

Google has upgraded its Gemini model to version 3.7 Flash, bringing significant improvements to its cloud-based agent, Spark. The new upgrade focuses on making workflows more reliable and efficient, especially for tasks like consolidating files, drafting emails, and updating status documents across Google Workspace.

Skyrocketing performance is a key feature of Gemini 3.7 Flash, with the model scoring 30.4% on AutomationBench compared to its predecessor's 17%. This puts it ahead of other notable models like GPT-5.6 Terra and Claude Sonnet 5.

The upgrade also brings improved performance in document-heavy work, with Gemini 3.7 Flash reaching 34% on GDP.pdf compared to 22% for the previous version. Additionally, it excels in coding-focused benchmarks, scoring 43.6% on FrontierCode and surpassing other models like Claude Sonnet 5 and GPT-5.6 Terra.

Under introductory pricing through December 31, Gemini 3.7 Flash costs $0.75 per million input tokens and $3.75 per million output tokens. This pricing is more competitive than that of its competitors, such as Claude Sonnet 5 at $2 and $10, and GPT-5.6 Terra at $2 and $12.