Google launched Gemini 3.7 Flash, a lower-cost artificial intelligence model focused on software coding, web development and multi-step AI agent tasks as the company continues competing with Anthropic and OpenAI.
The new model is designed for businesses and developers building AI systems that can plan tasks, use software tools and complete workflows with less human intervention. Google positioned Gemini 3.7 Flash as its latest “workhorse” model for coding and agent-based applications.
Gemini 3.7 Flash arrived about three weeks after Gemini 3.6 Flash. Google said the newer model improves coding, debugging, issue resolution, first-pass code accuracy and production-ready code generation. It also targets web development, UI generation, instruction following and multi-step planning.
The model's pricing is one of the main changes. Google is offering Gemini 3.7 Flash at an introductory rate of $0.75 per million input tokens and $3.75 per million output tokens through the end of 2026. From January 2027, the rates are scheduled to increase to $1.50 and $7.50 respectively, according to Google's launch information.
The pricing puts the model in a position aimed at high-volume AI workloads, where companies need to run large numbers of automated tasks without paying the cost associated with higher-end models for every request.
Google is also connecting Gemini 3.7 Flash to its broader AI agent ecosystem. The model is available through the Gemini API, Google AI Studio, Android Studio and Google Antigravity, while enterprise customers can access it through Google's enterprise agent products.
For individual users, Gemini Spark, Google's personal AI agent, is also being powered by Gemini 3.7 Flash for Google AI Pro and Ultra subscribers in supported markets. Google says Spark can use Workspace tools such as Gmail, Calendar and Docs while handling multi-step tasks.
The coding improvements are particularly important because AI-assisted software development has become one of the most competitive areas of the generative AI market. Google's latest model is aimed at reducing the number of prompts and manual corrections required to complete software-development tasks.
Independent user testing has also focused heavily on coding and agent workflows since the launch. Early community feedback has been positive in several cases, although those reports are anecdotal and should not be treated as controlled benchmarks.
Google's release also comes amid questions about the timing of its premium Gemini models. Reuters reported that investors have been watching Gemini 3.5 Pro as a measure of whether Google's DeepMind can keep pace with Anthropic and OpenAI. Google had previously said the model was being tested with partners and would arrive “soon”, but the company did not provide a launch date alongside Gemini 3.7 Flash.
The timing adds significance to the Flash release. Google is pushing a faster and cheaper model into commercial workflows while the market continues to wait for its next premium offering.
Google's AI strategy has also undergone organisational changes. Reuters reported that DeepMind chief Demis Hassabis stepped aside in favour of his deputy Koray Kavukcuoglu, while two original technical co-leads of Gemini left to establish a startup.
The company has continued to defend its position in the AI race. Chief executive Sundar Pichai has argued that Google's infrastructure, models and distribution across its products give the company a strong position despite intense competition.
For developers, the immediate attraction of Gemini 3.7 Flash is the combination of coding capability, agent workflows and lower introductory pricing. For Google, the larger test will be whether developers and businesses adopt the model at scale while the company prepares its next premium Gemini release.
Filed by
Startup Unplugged
