Google has launched Gemini 4 Argon, the first model in its Gemini 4 generation, introducing it with access initially limited to trusted cybersecurity defenders and Google's internal teams.

The company announced Argon on Sept. 30 through its Fairwind Programme, an initiative that gives selected governments and trusted partners access to Google's advanced cybersecurity models. According to the official announcement, Argon can autonomously find, validate and patch critical software vulnerabilities. Google is withholding wider access for now while it continues testing the model and refining its safeguards.

Paid application programming interface (API) customers and Google AI Ultra subscribers will gain access later, though Google has not announced a specific date for the broader rollout.

How Google says Argon compares

Google reports that Argon scored 77.9% on DeepSWE v1.1, a benchmark for software engineering tasks. On the same test, the company reports 74.2% for Anthropic's Claude Opus 5.5 and 74.1% for OpenAI's GPT-6 Astra.

On CWE-bench v1, which evaluates how well AI models identify and fix software vulnerabilities, Google reports that Argon scored 68%, tying GPT-6 Astra. The benchmark figures come from Google's own evaluations, so they do not independently establish how Argon performs across real-world workloads.

Argon can also generate up to 1 million tokens in a single response, compared with a 64,000-token output limit on previous models. Tokens are the basic units of text processed and generated by AI models. The larger output limit is designed to let Argon handle longer software engineering and other complex tasks without repeatedly being prompted to continue.

Internal deployment

Google says Argon is already in use internally by thousands of employees for coding, research and other specialized tasks. One reported use is migrating C and C++ codebases to Rust, a systems programming language designed to prevent the memory-safety errors that are a well-documented source of software vulnerabilities. According to Google, Argon agents are working on codebases ranging from tens of thousands of lines to more than 800,000 lines in the Zircon kernel of its Fuchsia operating system.

The company says the larger code migrations undergo automated and manual audits, emulation testing and review before being deployed to production. Kernel code sits at the core of an operating system, where errors can affect every component running on top it.

Google has also used Argon agents to analyze data-center performance data and identify memory optimizations. The company says the changes are expected to free more than 300 tebibytes of memory once fully rolled out, with total savings estimated at between 500 tebibytes and 1 pebibyte. A pebibyte is equivalent to 1,024 tebibytes.

Google's quantum computing researchers are using Argon to optimize algorithms as well. In one example, the company said Argon beat a published baseline for a quantum computing task by 40% in a matter of minutes.

Why access is being phased

Argon's cybersecurity capabilities are a major reason for the phased rollout. Google says the model can identify and fix vulnerabilities with limited human intervention, creating both defensive value and potential security risks as such capabilities become more widely available. The company is therefore making the model available first to trusted cyber defenders through Fairwind while it gathers feedback and continues developing safeguards.

Google said Argon will eventually be made available to developers, enterprises and consumers as access expands.

Pricing

Google will launch Argon at an introductory price of $2 per million input tokens and $10 per million output tokens. After the introductory period, the price will increase to $4 per million input tokens and $20 per million output tokens. Cached input tokens, which allow the same context to be reused across requests, will receive a 95% discount on the input-token price.

The launch comes as Google competes with OpenAI and Anthropic in the development of increasingly capable AI models for coding, cybersecurity, research and enterprise work. For now, however, Google's own benchmarks remain the main publicly available performance measurements for Argon; wider access will allow independent users and researchers to test the model across more workloads.