Google DeepMind Launches Gemini 3.7 Flash, Doubling Down on Coding and Agents

Google DeepMind today announced Gemini 3.7 Flash, the latest in its Flash series and its most intelligent workhorse model yet for coding and agents, just three weeks after the release of Gemini 3.6 Flash. The model shows strong gains over its predecessor in coding tasks, achieving higher first-pass code accuracy and better production-ready code generation, as evidenced by scores on FrontierCode 1.1 Main (43.6% vs 34.4%) and DeepSWE v1.1 (65.3% vs 49.0%). In web development, it outperforms 3.6 Flash on WebDev Arena with an Elo score of 1588 vs 1538, and it significantly improves on complex document processing in the GDP.pdf benchmark (34.0% vs 22.0%) and real-world business workflows in AutomationBench (30.4% vs 17.0%). The model is available at an introductory price of $0.75 per million input tokens and $3.75 per million output tokens through the end of the year. It also powers Gemini Spark, a personal AI agent for Google AI Pro and Ultra subscribers, enhancing tool use for Google Workspace apps. Safety updates include stronger safeguards against misuse in chemical, biological, radiological, and nuclear (CBRN) domains and cyber offense. Developers can access the model via Google Antigravity, the Gemini API through Google AI Studio and Android Studio, while enterprises can use it on the Gemini Enterprise Agent Platform.