Google drops Gemini 3.7 Flash model, and it’s ready to handle your chores with the Spark agent

Google has released Gemini 3.7 Flash, an upgraded version of its AI model, which is now being utilized by the Gemini Spark agent. Spark, a cloud-based agent, benefits from improved tool use and enhanced performance in multi-step tasks, making it capable of performing chores such as consolidating files, drafting emails, and updating status documents across various Google Workspace services like Gmail, Drive, Docs, Calendar, Keep, and Tasks.

The agent can operate even when a laptop is closed or a phone is locked and can navigate websites using saved accounts and browser passwords, handling tasks like comparing flight options and scheduling apartment viewings. Despite this advanced functionality, Google ensures that sensitive actions like payments are handled with caution.

Gemini 3.7 Flash is designed to produce more reliable workflows, as evidenced by its improved planning and tool call capabilities. The model's performance has been bolstered, scoring 30.4% on AutomationBench, outpacing GPT-5.6 Terra at 23.6% and Claude Sonnet 5 at 10.7%. Moreover, Gemini 3.7 Flash excels in document-heavy tasks, achieving a 34% score on GDP.pdf, surpassing other models' performances.

This upgrade comes at a lower cost, with $0.75 per million input tokens and $3.75 per million output tokens for introductory pricing through December 31, making it more affordable than its competitors. While Google's more advanced Gemini 3.5 Pro model is still forthcoming, the introduction of Gemini 3.7 Flash and the merging of Microsoft's Copilot apps into a unified experience represent significant advancements in the AI landscape, promising more efficient and reliable AI assistance.