Google released Gemini 3.7 Flash on Thursday, calling it "our most intelligent workhorse model yet for coding and agents," arriving just three weeks after Gemini 3.6 Flash, the company said.
On FrontierCode 1.1 Main, a production code quality benchmark, the model scored 43.6%, up from 34.4% for 3.6 Flash, Google said. On DeepSWE v1.1, a long-horizon software engineering benchmark, it scored 65.3%, up from 49.0%, and its WebDev Arena Elo score rose to 1588 from 1538.
The model also improved on GDP.pdf, a complex document processing benchmark, to 34.0% from 22.0%, and on AutomationBench, a business workflow automation benchmark, to 30.4% from 17.0%, according to Google.
Gemini 3.7 Flash is available now to developers through Google AI Studio, Android Studio and Google Antigravity, to enterprises through the Gemini Enterprise Agent Platform, and to consumers through Gemini Spark for Google AI Pro and Ultra subscribers in more than 160 countries, Google said. Introductory pricing runs $0.75 per million input tokens and $3.75 per million output tokens through Dec. 31, rising to $1.50 and $7.50 after that. Developer Simon Willison added support for the model to his llm-gemini plugin the same day, along with two new embedding models, gemini-embedding-2 and gemini-embedding-001, according to his site.
For developers, a coding-focused Flash model at introductory pricing through year end is a cost lever worth testing now, before the per-token price roughly doubles in January.