Google has rolled out Gemini 3.7 Flash, positioning it as the company's most capable workhorse model yet for coding and AI agent tasks, even as it continues to withhold a release date for its long-delayed flagship model, Gemini 3.5 Pro.
New model targets coding and agent workflows
The company announced Gemini 3.7 Flash on August 13, describing it as a direct outcome of developer feedback and internal algorithmic improvements. The release comes just three weeks after Gemini 3.6 Flash, continuing Google's rapid cadence of updates within the Flash family.
According to Google, the new model shows marked gains over its predecessor in software engineering tasks such as debugging and issue resolution, along with improved first-pass accuracy when generating production-ready code. On the FrontierCode 1.1 Main benchmark, the model scored 43.6 per cent compared to 34.4 per cent for Gemini 3.6 Flash, while on DeepSWE v1.1 it recorded 65.3 per cent against 49.0 per cent for the older model.
In web development, Google said Gemini 3.7 Flash produces more functional layouts and feature-complete applications using fewer prompts, and demonstrates stronger design adherence when working from reference inputs such as screenshots or full design systems. On Arena.ai's WebDev Arena leaderboard, the model posted an Elo score of 1588, up from 1538 for Gemini 3.6 Flash.
The company also highlighted improvements in knowledge-intensive domains including finance, law and biosciences. On the GDP.pdf benchmark, which tests a model's ability to process complex documents, Gemini 3.7 Flash scored 34.0 per cent against 22.0 per cent for its predecessor. On AutomationBench, which evaluates real-world business workflow completion, the new model scored 30.4 per cent compared to 17.0 per cent previously.
Gemini 3.7 Flash: Pricing and availability
Google is offering Gemini 3.7 Flash at an introductory price of $0.75 per million input tokens and $3.75 per million output tokens, roughly half the cost of Gemini 3.6 Flash. The company said this pricing will hold through the end of 2026, after which rates will rise to $1.50 per million input tokens and $7.50 per million output tokens from January 1, 2027.
The model is available to developers through Google Antigravity, Google AI Studio and Android Studio, while enterprises can access it via the Gemini Enterprise Agent Platform and the Gemini Enterprise app. For individual users, Gemini 3.7 Flash now powers Spark, Google's round-the-clock AI agent available to Google AI Pro and Ultra subscribers in more than 160 countries. Google said the update improves Spark's tool integration with Workspace applications, enhancing accuracy for multi-step tasks such as consolidating files, drafting emails and updating status documents.
The company added that Gemini 3.7 Flash ships with updated safeguards against misuse in chemical, biological, radiological and nuclear domains, as well as cyber offense, in line with its existing safety frameworks.
Gemini 3.7 Flash: Flagship model remains in limbo
Despite the steady stream of Flash releases, Google has not disclosed a launch timeline for Gemini 3.5 Pro, its most powerful model, which has faced repeated delays. Bloomberg reported that the delay has fuelled investor concerns about Google's product roadmap, particularly in AI coding, an area seen as commercially critical.
Alphabet CEO Sundar Pichai said during the company's July earnings call that Google intended to release models at a faster pace, and confirmed that significant computing resources were already being directed toward training the upcoming Gemini 4 model. Bloomberg had earlier reported that Gemini 3.5 Pro's delay stemmed from an internal push to strengthen the model's coding capabilities, with Google reorganising engineering teams within DeepMind to consolidate its AI coding efforts.
