# Google announces new Gemini 3.8 Flash model More capable model may come with higher costs - **Source:** [The Verge (theverge.com)](https://www.theverge.com) - **Published:** 2026-09-02 - **Topic:** [Data & AI](https://brifing.si/c/data-ai) - **Original article:** - **On Brifing:** https://brifing.si/s/6f01eb2d-90ab-4bb2-afa0-a88a01704de3 ## Key features of the new model Google has unveiled the Gemini 3.8 Flash model. The new model performs more reasoning steps on complex tasks and calls tools iteratively. The company claims the model works harder than its predecessor. While introductory pricing remains the same as 3.7 Flash, costs for users may increase. Google warns that the model might use more tokens to maximize performance, especially at higher effort levels. ## Pricing and performance comparison Gemini 3.8 Flash is priced at $0.75 per million input tokens and $3.75 per million output tokens. Artificial Analysis notes it's the cheapest option at this intelligence level. However, costs could rise up to 40% due to a 30% increase in output tokens per task and more turns on agentic evaluations. Aigora.ai CEO John Ennis highlights the model's cost-effectiveness and speed compared to Anthropic's models. ## Performance in software engineering Google announced significant improvements in software engineering and autonomous AI agents with Gemini 3.8 Flash. The model outperformed its predecessor and other frontier models on the DeepSWE v1.1 benchmark. While Anthropic's Fable 5 also received an update this week, Google's new model stands out in price-performance ratio. Costs were further reduced by lowering prices for cached data usage. ## Success in finance and legal domains Gemini 3.8 Flash also outperformed competitors on the Vals Finance Agent V2 and Harvey's Legal Agent benchmarks. The model shows superior performance in financial and legal tasks. However, Google stated the new model includes safeguards against misuse in Chemical, Biological, Radiological, and Nuclear (CBRN) domains and cyber offenses. This indicates the model is designed to minimize security vulnerabilities. ## Fairwind Program and 3.8 Flash Cyber Google launched Gemini 3.8 Flash Cyber alongside its new Fairwind Program, limited to governments and trusted partners. The program's 650 members include organizations like CrowdStrike and Center for Internet Security. It provides access to 3.8 Flash Cyber and Google's CodeMender agent, which can autonomously detect and fix vulnerabilities to protect critical infrastructure, public services, and national security. ## Availability and subscription options Gemini 3.8 Flash is now available for consumers with Google AI Pro or Ultra subscriptions, as well as developers and enterprise users. The model's wide availability could contribute to the democratization of AI technologies. Google aims to expand its market share by offering various subscription options catering to different user groups. This strategy may strengthen the company's leadership in the AI space. ## Industry impact The launch of Gemini 3.8 Flash is seen as a significant step in the AI industry. Its balance of performance and cost may push competitors to act. The advantages it offers in fields like software engineering and finance could enhance companies' operational efficiency. Google's security measures set an example for responsible use of AI technologies. ## Future projections Google's new model provides insights into the future of AI technologies. The development of smarter and more efficient models seems likely to intensify industry competition. Initiatives like the Fairwind Program could ensure AI is used safely in critical areas. Google's steps may establish new standards in AI ethics and security. ## Sources covering this story - https://www.theverge.com/ai-artificial-intelligence/988742/google-gemini-3-8-flash — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._