Google Releases New Gemini 3.7 Flash Model
Google has released a new AI model, Gemini 3.7 Flash, just three weeks after its previous Flash version update. The new model aims to improve coding capabilities and complex document processing.
Google has announced the release of its Gemini 3.7 Flash AI model, replacing the recently updated 3.6 Flash version. This marks another rapid iteration, with the new model rolling out only three weeks after its predecessor.
The Gemini 3.7 Flash model is designed to offer enhanced performance, particularly in coding and agentic tasks. Google reports significant gains in coding benchmarks, with the FrontierCode 1.1 Main test score increasing from 34.4 percent to 43.6 percent, and the DeepSWE v1.1 score rising from 49 percent to 65.3 percent.
Improvements are also noted in the model's ability to process complex documents. The GDP.pdf benchmark, which assesses document comprehension, saw a rise to 34 percent from 22 percent with the previous 3.6 Flash version. Furthermore, AutomationBench tests, measuring the execution of business workflows, showed an increase from 17 percent to 30.4 percent.
Google is also positioning Gemini 3.7 Flash with a competitive introductory price, aiming to address the cost concerns associated with some alternative AI models.