Introducing Gemini 3.7 Flash

Google has announced the release of Gemini 3.7 Flash, an updated version of their widely used Flash series of models. This latest iteration, introduced just three weeks after Gemini 3.6 Flash, comes in response to developer feedback and algorithmic advancements. The new model offers significant improvements across various fields of work, including software engineering, knowledge work, and web development, and is priced at half the cost of the original 3.6 Flash per million tokens.

3.7 Flash demonstrates substantial gains in coding tasks such as debugging and issue resolution, with higher first-pass code accuracy and improved production-ready code generation. In web development, the model creates more functional layouts and feature-complete applications in fewer prompts, outperforming its predecessor on Arena.ai's WebDev Arena.

For knowledge-dense fields like finance, law, and biosciences, 3.7 Flash shows enhanced reasoning and accuracy, surpassing 3.6 Flash on benchmarks such as GDP.pdf and AutomationBench.

The update also brings about a more improved developer experience, with the model better adapting to challenges, clarifying intent, and following instructions more faithfully. It also thinks more diligently, investing more effort into multi-step planning and tool calls, thereby reducing the need for manual oversight and retries.

Starting today, Gemini Spark, a personal AI agent available to Google AI Pro and Ultra subscribers in over 160 countries, will utilize Gemini 3.7 Flash. With this update, Spark becomes more efficient for knowledge work, particularly in Google Workspace apps, delivering more accurate and high-quality outputs for complex, multi-skill workflows. For instance, it can consolidate files, draft emails, and update status documents more effectively than before.

Google also continues to improve the safety and robustness of its Frontier Safety safeguards. Gemini 3.7 Flash includes updated defenses against misuse in sensitive domains like Chemical, Biological, Radiological, and Nuclear (CBRN) and cyber offense, while still enabling beneficial use cases in line with Google's bioresilience and cyber program. More information on these updates can be found on the 3.7 Flash model card.

The introductory pricing for 3.7 Flash will expire on December 31, 2026, after which the cost will increase to $1.50/1M input tokens and $7.50/1M output tokens.