Google has launched Gemini 3.7 Flash, its newest workhorse AI model built for coding, web development, knowledge work, and agent-based business tasks.
The model was announced on August 13, 2026, only three weeks after Gemini 3.6 Flash, making it another fast follow-up in Google’s Flash lineup.
Reuters reported that Google is positioning the model as a lower-cost option for businesses building autonomous AI systems that can plan tasks, use software tools, and complete multi-step workflows with less human help.
For developers, the price may be the bigger headline.
Google is offering Gemini 3.7 Flash 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. Reuters reported that the pricing is half the original cost of Gemini 3.6 Flash.
That matters because AI agents are getting more expensive to run.
Unlike simple chatbots, agents often need to read files, plan steps, call tools, write code, test outputs, and revise their work. A cheaper Flash model gives developers more room to build apps that run several model calls inside one workflow.
Gemini 3.7 Flash focuses on coding and agents
Google says Gemini 3.7 Flash improves on Gemini 3.6 Flash in software engineering tasks, including debugging, issue resolution, and production-ready code generation. Reuters also reported that the model is aimed at automated business workflows and AI systems that can complete tasks with less manual intervention.
Quartz, citing Google’s figures, reported that Gemini 3.7 Flash scored 43.6 percent on FrontierCode 1.1, up from 34.4 percent for Gemini 3.6 Flash. It also scored 65.3 percent on DeepSWE v1.1, compared with 49.0 percent for the previous Flash model.
The model also appears to be stronger in web development. Quartz reported that Google highlighted gains in generating more functional layouts, following design instructions, and producing feature-complete apps with fewer prompts.
That gives Google a practical pitch: not just a smarter model, but one that can reduce retries.
For coding agents, fewer failed attempts can mean lower cost, faster completion, and less developer supervision.
Google also targets business automation
Gemini 3.7 Flash is not only for code.
Google is also pitching the model for knowledge-heavy work in fields such as finance, law, and biosciences. Quartz reported that the model improved on GDP.pdf, a benchmark for complex document processing, and AutomationBench, which tests real-world business workflow completion.
That puts Gemini 3.7 Flash in the middle of a larger shift in AI tools.
Companies are no longer asking models only to answer questions. They want models to read reports, compare documents, update spreadsheets, draft emails, search files, and complete longer processes across several apps.
That is where pricing becomes important again.
A model that is good enough for repeat agent calls, but cheaper than larger frontier models, could become more attractive for teams building everyday workplace automations.
Gemini Spark gets the new model
Google is also rolling Gemini 3.7 Flash into Gemini Spark, its personal AI agent for Google AI Pro and Ultra subscribers.
Reuters reported that Gemini Spark is available to Google AI Pro and Ultra customers in more than 160 countries. Google AI Studio also said on X that Spark would start using Gemini 3.7 Flash for subscribers in over 160 countries.
With the upgrade, Spark is expected to handle more complex tasks across Google Workspace apps, including Gmail, Google Calendar, and Google Docs.
That may be one of Google’s biggest advantages in the AI agent race. The company already controls many of the apps where people work every day. If Gemini can reliably act across those tools, Google does not need to sell agents as a separate experience. It can build them directly into the workflow.
Developers can access Gemini 3.7 Flash now
Gemini 3.7 Flash is available through the Gemini API, Google AI Studio, Android Studio, Google Antigravity, and Google’s enterprise agent tools, according to Google-related announcements shared online and covered by Reuters and Quartz.
Google DeepMind CEO Demis Hassabis also highlighted the model on X, saying Gemini 3.7 Flash brings major upgrades for software engineering, web development, and knowledge work.
Sundar Pichai also promoted the launch on X, while Logan Kilpatrick noted the model’s lower price and faster performance.
The release comes as Google faces pressure to keep pace with OpenAI and Anthropic in coding, agents, and frontier AI models. Reuters noted that Google did not give a release date for Gemini 3.5 Pro, the larger model investors have been watching as a sign of DeepMind’s progress.
For now, Gemini 3.7 Flash looks like a practical developer release rather than a flashy consumer update.
The pitch is simple: better coding, stronger agents, wider availability, and lower temporary pricing.
That may be enough to make the model interesting to developers building AI-powered apps, workplace tools, and business automation systems at scale.
As of this writing, Gemini 3.7 Flash is available across Google’s developer, enterprise, and Gemini Spark channels.
via: Google Blog | Google DeepMind | Reuters | Quartz | X / Demis Hassabis
