Google has launched Gemini 3.7 Flash, its latest Flash-series AI model focused on coding, software engineering, web development and AI agent workflows. Google calls it its most intelligent “workhorse” model yet for coding and agents. The model arrives just three weeks after Gemini 3.6 Flash and brings improvements in reasoning, instruction following, tool use and multi-step task execution.
Gemini 3.7 Flash: What’s New?
Gemini 3.7 Flash is designed for complex tasks where AI needs to plan, use tools, solve problems and adapt to obstacles rather than simply generate an answer.
Google says the model delivers improvements in:
- Coding and debugging
- Software engineering
- Web and UI development
- AI agent workflows
- Business automation
- Complex document analysis
- Finance, law and bioscience-related knowledge work
The model can also put more effort into multi-step planning and tool calls, potentially reducing retries and the amount of manual supervision required.
Better Coding Performance
Coding is one of the biggest areas of improvement in Gemini 3.7 Flash.
Google reports a 43.6% score on FrontierCode 1.1 Main, compared with 34.4% for Gemini 3.6 Flash. On DeepSWE v1.1, the new model scored 65.3% versus 49.0% for its predecessor.
The model is designed to improve debugging, issue resolution, first-pass code accuracy and production-ready code generation.
Stronger AI Agent Capabilities
Gemini 3.7 Flash is also built for agentic workflows, where an AI system can break a larger objective into multiple steps and use external tools to complete it.
Google says the model improved significantly on AutomationBench, scoring 30.4% compared with 17.0% for Gemini 3.6 Flash. This makes it suitable for applications such as coding agents, research assistants and business-process automation.
Improved Web Development
The new model can also generate more functional websites and feature-complete applications using fewer prompts.
On WebDev Arena, Gemini 3.7 Flash achieved an Elo score of 1588, compared with 1538 for Gemini 3.6 Flash. Google says it can also better follow reference screenshots, images and complete design systems when generating user interfaces.
Better Complex Document Understanding
Google has improved Gemini 3.7 Flash for knowledge-heavy tasks as well.
On the GDP.pdf benchmark, which tests complex document processing, the model scored 34.0% compared with 22.0% for Gemini 3.6 Flash. Google highlights potential applications in areas such as finance, law and biosciences.
Gemini 3.7 Flash Price
Google has announced introductory API pricing of:
- $0.75 per 1 million input tokens
- $3.75 per 1 million output tokens
This introductory pricing is available through the end of 2026. The combination of lower cost and improved performance is aimed at making large-scale AI agent deployment more economical.
Where Is Gemini 3.7 Flash Available?
Developers can access Gemini 3.7 Flash through Google AI Studio, Gemini API, Google Antigravity and Android Studio. Enterprise customers can also use it through Google’s Gemini Enterprise platforms.
Google is also bringing Gemini 3.7 Flash to Gemini Spark for Google AI Pro and Ultra subscribers in more than 160 countries. Spark can use the model for tasks such as working with files, drafting emails and updating documents.
Final Verdict
Gemini 3.7 Flash is primarily designed to make AI better at completing complicated tasks rather than simply answering questions. Its stronger coding, reasoning, tool-use and agent capabilities could make it a useful option for developers building autonomous AI applications.
With improved benchmark performance and introductory pricing of $0.75/$3.75 per million input/output tokens, Gemini 3.7 Flash is positioned as a faster and more cost-efficient option for production AI agents.
About the Author
