Faster and cheaper: Gemini 3.7 Flash targets coding and agents

Gemini 3.7 Flash improves coding, agents, and complex workflows, with gains across several benchmarks and introductory pricing temporarily cut in half.

Three weeks after Gemini 3.6 Flash, Google is expanding the lineup with Gemini 3.7 Flash, a model focused on software development, agents, and knowledge-work workflows. The new version improves instruction following, multi-step planning, and tool use, with the aim of reducing manual intervention on complex tasks.

The most visible gains are in coding. Google reports a score of 65.3% on DeepSWE v1.1, up from 49% for 3.6 Flash, and 43.6% on FrontierCode 1.1 Main, compared with 34.4%. On WebDev Arena, Gemini 3.7 Flash reaches an Elo score of 1,588, versus 1,538 for its predecessor. The company also highlights better adherence to visual references when generating interfaces from a screenshot, image, or design system.

The model also improves on document-heavy and professional tasks. On GDP.pdf, which focuses on processing complex documents, the reported score rises from 22% to 34%. AutomationBench, which evaluates the execution of real-world business workflows, increases from 17% to 30.4%. Google also reports better performance in knowledge-intensive fields such as finance, law, and biosciences.

Part of this improvement comes from more sustained behavior during planning and tool calls. In Antigravity, Gemini 3.7 Flash can orchestrate sub-agents, use MCP servers, and pursue goals across long-running trajectories. Google demonstrates use cases involving interface generation, cloud data analysis, web experience creation, and multi-agent loops applied to robotics.

Gemini Spark is also moving to 3.7 Flash. Google’s personal agent gains improved use of Workspace tools for consolidating files, drafting emails, and updating status documents within workflows that combine multiple skills.

Pricing is the other notable change. Through the end of the year, Gemini 3.7 Flash is priced at $0.75 per million input tokens and $3.75 per million output tokens, half the original price of Gemini 3.6 Flash. Starting in 2027, Google plans to raise pricing to $1.50 for input and $7.50 for output. The model is available through the Gemini API, Google AI Studio, Android Studio, Google Antigravity, Gemini Enterprise, and Gemini Spark.