Google has launched Gemini 3.7 Flash, a highly efficient model optimized for software engineering, complex agentic workflows, and knowledge-intensive reasoning. Released just three weeks after version 3.6 Flash, this new iteration focuses on improving multi-step planning, tool-calling accuracy, and instruction fidelity. Technical benchmarks highlight significant performance leaps: the model achieved a 43.6% score on FrontierCode 1.1 Main compared to 34.4% for its predecessor, and reached 65.3% on DeepSWE v1.1 against 49.0%. In web development, it reached an Elo score of 1588 on Arena.ai’s WebDev Arena. Furthermore, the model demonstrates superior document processing capabilities, scoring 34.0% on the GDP.pdf benchmark versus 22.0% for 3.6 Flash, and nearly doubling its performance on AutomationBench (30.4% vs 17.0%). For developers, the model is offered at an introductory price of $0.75 per million input tokens and $3.75 per million output tokens, representing a 50% cost reduction compared to the previous version. Additionally, Gemini 3.7 Flash is being integrated into Gemini Spark to enhance its ability to execute multi-skill workflows across Google Workspace. The deployment includes updated Frontier Safety safeguards specifically targeting CBRN and cyber offense domains.