--- title: "Gemini 4 Argon goes to cyber defenders first, at $2 per million" url: https://blog.vomini.com/blog/gemini-4-argon-announced date: 2026-10-02T13:30:00+00:00 updated: 2026-10-27T22:42:30.621344+00:00 author: "Roberto R." section: "News" language: en site: "Vomini" license: free to quote with attribution --- # Gemini 4 Argon goes to cyber defenders first, at $2 per million > Google announced Gemini 4 Argon, its new frontier model, with an introductory price of $2 per million input tokens and $10 per million output tokens. It starts with vetted cyber defenders, without cyber guardrails for… On 30 September, Google announced Gemini 4 Argon, its new frontier model. It is not open to everyone yet. Argon is rolling out first to a group of trusted cyber defenders through Google's Fairwind Program, and Google says it is taking part in the US government's voluntary process for testing models before release. The price is already set. Argon will launch at an introductory $2 per million input tokens and $10 per million output tokens, with cached input 95% cheaper than regular input. ## Why cyber defenders get it first Google says it trained Argon to find, check and patch serious software vulnerabilities on its own. For trusted defenders and Google's internal teams, Argon is being released without cyber guardrails, so they can use the full capability. Google cites Wiz, a cloud security company, which it says used Argon to find a critical vulnerability exposing sensitive personal data in healthcare software used by hospitals. On CWE-bench v1, a test of fixing security flaws, Google says Argon ties for first place at 68%. For everyone else, the model is designed to refuse requests that could help cyber or weapons attacks. Google also says it will monitor Argon's reasoning and actions and stop execution when the model goes beyond what the user intended. ## What else Google claims Google reports 77.9% on DeepSWE v1.1, a test of long software engineering tasks, and 51.3% on Zapier's AutomationBench, where it says Argon ranks first. The output limit grows to 1 million tokens, up from 64,000, which means the model can write very long answers or work through long problems in one go. Inside Google, the company says Argon agents freed more than 300 TiB of memory across its data centers and are porting large C and C++ codebases to Rust, with heavy review before anything reaches production. ## When the rest of us get it Google says broad access will start with paid API customers and Google AI Ultra subscribers, after more feedback from early testers. The post gives no date and does not say what the price becomes after the introductory period. ## What to do now If you build on the Gemini API, there is nothing to switch yet. Use the time to estimate what your current workload would cost at $2 and $10 per million tokens, with cached input at 95% off, and compare that with your current bill. If you work in security, look at whether your organization qualifies for the Fairwind Program. ## What we think Releasing a frontier model to defenders before attackers can buy it is a sensible order. The announced price is the other thing to note: at $2 and $10 it sits well below the $10 and $50 OpenAI charges for [GPT-6 Astra](/blog/gpt-6-astra-launch). If that introductory price holds once Argon is broadly available, it will put pressure on what the other labs charge. ## Sources - [Google](https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-4-argon/), 30 Sep 2026 - [OpenAI, GPT-6 Astra announcement (for the Astra price)](https://openai.com/index/gpt-6-astra/), 3 Sep 2026 --- Originally published at https://blog.vomini.com/blog/gemini-4-argon-announced