Google has introduced Gemini 3.7 Flash, a new version of its Flash series that the company says is designed to handle coding, software development and AI agent workflows more effectively while costing less to run.
The model arrives just three weeks after Gemini 3.6 Flash and is being marketed as a faster, more capable option for developers working on complex, multi-step tasks. Google says improvements in the new model came from developer feedback as well as changes to its underlying algorithms.
Coding is one of the areas where Google claims the biggest gains. Gemini 3.7 Flash is said to be better at debugging, resolving issues and producing usable code on its first attempt. On FrontierCode 1.1 Main, it recorded a score of 43.6 per cent, compared with 34.4 per cent for Gemini 3.6 Flash. On DeepSWE v1.1, the newer model scored 65.3 per cent, up from 49 per cent.
Google also reports improvements in web development. Gemini 3.7 Flash can generate functional layouts and more complete applications with fewer prompts, while also following visual references more closely. These references can include screenshots, images or an entire design system.
On Arena.ai’s WebDev Arena, Gemini 3.7 Flash achieved an Elo score of 1,588, compared with 1,538 for its predecessor.
The company is also targeting areas where AI systems need to process large amounts of information and follow multiple steps. Google says the new model performs better in knowledge-heavy areas such as finance, law and biosciences.
On the GDP.pdf benchmark, which tests how models handle complex documents, Gemini 3.7 Flash scored 34 per cent, compared with 22 per cent for Gemini 3.6 Flash. It also recorded a higher score on AutomationBench, reaching 30.4 per cent versus 17 per cent for the previous model. The benchmark is designed around completing real-world business workflows.
Google says the model has also been tuned to require less supervision. It can respond to roadblocks, ask for clarification when necessary and follow instructions more closely. The company says it puts greater effort into multi-step planning and tool calls, potentially reducing retries and manual intervention during development work.
Cost is another major change. Gemini 3.7 Flash will be available at an introductory price of $0.75 per million input tokens and $3.75 per million output tokens through the end of the year. Google says the pricing is half the original cost of Gemini 3.6 Flash per million tokens, making the new model cheaper to deploy at scale.
The update is also being applied to Gemini Spark, Google’s personal AI agent. Spark, available to Google AI Pro and Ultra subscribers in more than 160 countries, will use Gemini 3.7 Flash from launch. Google says the upgrade should improve Spark’s handling of Google Workspace tools and complex workflows involving multiple skills.
For developers, Gemini 3.7 Flash is available through Google AI Studio and the Gemini API, as well as Android Studio and Google Antigravity. Enterprise users can access it through the Gemini Enterprise Agent Platform and Gemini Enterprise app.
Google has also highlighted safety changes accompanying the release. Gemini 3.7 Flash includes updated safeguards covering potential misuse involving chemical, biological, radiological and nuclear applications, as well as cyber offences. The company said the measures are intended to restrict harmful applications while allowing beneficial uses.
Disclaimer: This content has not been generated, created or edited by Finance SC. Publisher:
Source link
Source link
