Google launches Gemini 3.7 Flash, claims major leap in coding and AI agents

Google has introduced the Gemini 3.7 Flash large language model (LLM), a new version of its Flash series that the company describes as its most capable "workhorse" model yet for coding and AI agents. The model arrives just three weeks after Gemini 3.6 Flash and is designed to improve software engineering, web development, and other complex workflows.

According to Google, Gemini 3.7 Flash delivers significant improvements over its predecessor in debugging, issue resolution, code generation, and multi-step tasks. The company says the model achieved a 43.6 % score on FrontierCode 1.1 Main, compared with 34.4 % for Gemini 3.6 Flash, while its score on DeepSWE v1.1 increased from 49 % to 65.3 %.

The new model also performs better at web development. Google says Gemini 3.7 Flash can generate more functional layouts and feature-complete applications using fewer prompts. It also shows improved adherence to reference designs, whether they are provided as screenshots, images, or complete design systems. On Arena.ai's WebDev Arena, the model achieved an Elo score of 1588, compared with 1538 for Gemini 3.6 Flash.

Google is also positioning Gemini 3.7 Flash for knowledge-intensive tasks in areas such as finance, law, and biosciences. On the GDP.pdf benchmark, which evaluates the ability to process complex documents, the new model scored 34 % versus 22 % for its predecessor. On AutomationBench, which measures performance on real-world business workflows, Gemini 3.7 Flash achieved 30.4 % compared with 17 % for Gemini 3.6 Flash.

The company says the model has also been designed to provide a better experience for developers. Gemini 3.7 Flash is reportedly better at adapting when it encounters obstacles, clarifying user intent, following instructions, planning multi-step tasks, and using external tools. Google says this should reduce the need for manual intervention and repeated attempts when building software agents.

One of the most notable changes is the pricing. Gemini 3.7 Flash is available at an introductory rate of $0.75 per 1 million input tokens and $3.75 per 1 million output tokens through the end of 2026. Google says this is half the original price of Gemini 3.6 Flash. From January 1, 2027, the price will increase to $1.50 per million input tokens and $7.50 per million output tokens.

Gemini 3.7 Flash is also being integrated into Gemini Spark, Google's 24/7 personal AI agent. Starting today, Spark will use the new model to handle knowledge-work tasks such as consolidating files, drafting emails, and updating status documents. Spark is available to Google AI Pro and Ultra subscribers in more than 160 countries.

For developers, Gemini 3.7 Flash is available through Google Antigravity, the Gemini API, Google AI Studio, and Android Studio. Enterprise customers can access it through Gemini Enterprise Agent Platform and the Gemini Enterprise app. Individual users can access the model through Spark in the Gemini app with Google AI Pro or Ultra subscriptions in supported countries.