Google has released Gemini 3.7 Flash, a new, cost‑effective workhorse model in the Flash series aimed at developers, agent orchestration, and knowledge‑intensive workflows. The company says the update follows directly on the heels of Gemini 3.6 Flash and reflects developer feedback and algorithmic improvements.
Improved performance in coding and knowledge work
According to Google, Gemini 3.7 Flash delivers notable gains over 3.6 Flash across several areas. On coding tasks such as debugging and issue resolution it achieves higher first‑pass code accuracy and better production‑ready code generation. For example, it scored 43.6% on the FrontierCode 1.1 Main benchmark versus 34.4% for 3.6 Flash, and 65.3% on DeepSWE v1.1 versus 49.0% for 3.6 Flash.
In web development, the model can produce functional layouts and feature‑complete applications in fewer prompts. For UI generation it shows strong design adherence to reference inputs — whether a screenshot, image, or full design system. On Arena.ai’s WebDev Arena the model achieved an Elo score of 1588 compared with 1538 for 3.6 Flash.
In knowledge‑dense domains such as finance, law, and biosciences, Google reports improved reasoning and accuracy. On the GDP.pdf benchmark — which tests the ability to process complex documents — Gemini 3.7 Flash scored 34.0% compared with 22.0% for 3.6 Flash. It also outperformed 3.6 Flash on AutomationBench (30.4% vs 17.0%), indicating stronger completion of real‑world business workflows.
Multimodal demos and agent use cases
Google highlighted a range of example applications: generating playable 3D game elements in real time when combined with Nano Banana; producing interactive landing pages in a single pass by orchestrating sub‑agents and using Gemini Omni for parallax components; a robotics training setup using a multimodal three‑agent graph to accelerate learning; and converting static PDFs into interactive data stories with live charts and aggregated insights.
Developer experience and pricing
Gemini 3.7 Flash is said to offer a noticeably improved developer experience compared with 3.6 Flash. The model adapts better to roadblocks, clarifies intent when necessary, follows instructions with higher fidelity, and engages more in multi‑step planning and tool calls. Google says this more disciplined execution reduces manual oversight and the number of retries in engineering workflows.
Introductory pricing is available through December 31, 2026: $0.75 per 1M input tokens and $3.75 per 1M output tokens. From January 1, 2027, the price will rise to $1.50 per 1M input tokens and $7.50 per 1M output tokens.
Spark and enterprise availability
Gemini Spark, which is available to Google AI Pro and Ultra subscribers in more than 160 countries, will start using Gemini 3.7 Flash immediately. Google launched Spark at I/O as a 24/7 personal AI agent that acts under user direction; the model update is intended to make Spark more efficient for knowledge work by improving tool use for Google Workspace apps and boosting accuracy and output quality for complex multi‑skill workflows.
Enterprises can access 3.7 Flash via the Gemini Enterprise Agent Platform and the Gemini Enterprise app. Developers can build with the model through the Gemini API in Google AI Studio and Android Studio; Google recommends the Google Antigravity demo environment for testing agent‑first workflows.
Safety and safeguards
Google states it continuously improves Frontier Safety safeguards. Gemini 3.7 Flash ships with updated protections to mitigate misuse in Chemical, Biological, Radiological, and Nuclear (CBRN) domains and against cyber‑offense, while enabling beneficial use cases in line with Google’s bioresilience and cyber programs. More details are available in the 3.7 Flash model card.
How to try it
- Developers: access the Gemini API via Google AI Studio or Android Studio and follow Google’s developer guide; test agent workflows in Google Antigravity.
- Enterprises: use the Gemini Enterprise Agent Platform and the Gemini Enterprise app.
- Individuals: access Gemini Spark in the Gemini app if you are a Google AI Pro or Ultra subscriber in a supported country.
Introductory pricing expires on December 31, 2026; new pricing becomes effective January 1, 2027.
