Google announced Gemini 3.7 Flash on August 13, calling it its “most intelligent workhorse model yet for coding and agents.” The release builds on the progress of Google’s widely used Flash series and arrives just three weeks after Gemini 3.6 Flash, a cadence Google attributes to developer feedback and algorithmic innovations it plans to carry into future models.
The model ships with an introductory price of $0.75 per million input tokens and $3.75 per million output tokens, about half the original per-token cost of Gemini 3.6 Flash, and that rate runs through the end of the year. Google says the combination of price and performance lets developers scale production-ready agents cost effectively. The discount also lands in a month of aggressive API pricing moves: OpenAI cut GPT-5.6 Luna pricing by 80 percent in early August, while DeepSeek signaled a major price hike of its own amid surging demand.
Benchmark gains over 3.6 Flash
Google reports strong gains on coding and issue-resolution tasks. On FrontierCode 1.1 Main, Gemini 3.7 Flash scores 43.6 percent against 34.4 percent for 3.6 Flash, and on DeepSWE v1.1 it reaches 65.3 percent, up from 49.0 percent. The company also highlights higher first-pass code accuracy and better results generating production-ready code, which is what FrontierCode measures.
In web development, Gemini 3.7 Flash posts an Elo of 1588 on Arena.ai’s WebDev Arena against 1538 for its predecessor, and Google says it generates more functional layouts and feature-complete apps in fewer prompts. For UI generation, the model shows high design adherence and parity based on a reference input, whether that reference is a screenshot, an image, or a full design system.
For knowledge work, the company cites gains on the GDP.pdf benchmark (34.0 percent versus 22.0 percent) and on AutomationBench (30.4 percent versus 17.0 percent), pointing to improved reasoning in finance, law, and biosciences. Google also describes a better developer experience: the model adapts to roadblocks, clarifies intent when needed, and follows instructions with greater fidelity, with more disciplined multi-step planning and tool calls that need less manual oversight.
Price, availability, and safety
Gemini 3.7 Flash is available to developers through the Gemini API in Google AI Studio and Android Studio, and in agent-first workflows in Google Antigravity. Enterprises can reach it through the Gemini Enterprise Agent Platform and the Gemini Enterprise app. For consumers, Gemini Spark, the 24/7 personal agent for Google AI Pro and Ultra subscribers in more than 160 countries, began using 3.7 Flash on the day of the announcement, consolidating files, drafting emails, and updating status documents across Google Workspace apps.
Google says early customer feedback highlights the model’s performance and precision at lower cost, and that the release ships with updated Frontier Safety safeguards against misuse in chemical, biological, radiological, and nuclear domains and in cyber offense, in line with its bioresilience and cyber programs. The same week brought the Pixel 11 launch with Google’s Tensor G6 chip, giving the company two major AI announcements in quick succession. The introductory API rate holds through the end of 2026.
