TL;DR
Get business pricing on tech for your team
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
Google announced Gemini 4 Argon on Sept. 30, 2026, and said it is rolling the model out to trusted cyber defenders through its Fairwind Program. Google reports strong results in software engineering, enterprise tasks and defensive cybersecurity, but has not published independent validation of the performance claims. Broader availability will follow phased testing and guardrail work.
Google announced Gemini 4 Argon on Sept. 30, 2026, and said it is beginning a limited rollout to trusted cyber defenders through its Fairwind Program. The company describes Argon as a frontier model for long, complex work in software engineering, enterprise fields such as legal and finance, and cybersecurity defense; it says broader access will depend on phased testing and guardrail development.
Google says Argon is already being used internally by thousands of its employees for specialized coding, research and writing. The company reported that a team of Argon agents identified memory optimizations in data-center profiling information that could free more than 300 tebibytes of memory once rolled out, with estimated total savings of 500 tebibytes to 1 pebibyte. Those are company estimates, and Google did not provide an independent assessment in the announcement.
Google also described a migration effort involving C and C++ codebases being rewritten in Rust, including work on the Fuchsia Zircon kernel. The company said those changes are undergoing automated and manual audits, emulation testing and review before production use. In a separate example, Google said Argon replaced 32,000 lines of SIMD code in a Rust port of its libgav1 video decoder; it reported that the resulting decoder ran 2.7 times faster than the prior Rust port with identical video output.
For external evaluations, Google reported a 77.9% score on DeepSWE v1.1, a benchmark for long-horizon software engineering, and a 51.3% score on AutomationBench, where it said Argon ranked first. Google also cited a 91.7% score on LVBench for long-video understanding and leading results on the Vals Index and selected legal and finance evaluations. These figures and rankings are reported by Google; the announcement does not establish independent verification or provide all comparison details.
A Restricted Start for Cyber Defense
The initial release puts the emphasis on controlled access to a model Google says can find, validate and patch software vulnerabilities. That capability could help defenders assess and fix flaws, while also making careful handling of advanced cybersecurity tools a central part of the release. Google says trusted defenders and its internal teams will receive Argon without cyber guardrails so they can use its full defensive capabilities.
For developers and businesses, the announcement signals Google’s intention to apply Argon to sustained, multi-step work rather than only short prompts. Google says the model’s output limit is expanding to 1 million tokens, up from 64,000, allowing it to generate much longer responses in a single task. The company has not yet opened general access, so the practical performance and availability for customers remain to be seen.
From Internal Use to Limited Testing
Google says Argon has been powering internal workflows as it prepares the model for outside use. Its announcement describes work on quantum algorithm optimization, memory use across data centers and large code migrations. Google said one quantum computing example beat a published baseline by 40% in minutes, but did not identify the baseline or provide further methodological detail in the supplied announcement.
The company says it is participating in the U.S. government’s voluntary pre-release model access process as it expands access in stages. It plans to gather feedback from early testers and iterate on safeguards before making Argon available to developers, enterprises and consumers. Google listed introductory API pricing at $2 per million input tokens and $10 per million output tokens, with cached input tokens priced 95% below the input rate.
““Today, we’re announcing our new frontier model, Gemini 4 Argon, which is rolling out to a set of trusted cyber defenders through our Fairwind Program.””
Independent Checks and Access Dates
The announcement does not give a date for general availability or specify how many testers are in the initial rollout. It also does not detail the scope of the cybersecurity guardrails being adjusted, or explain how Google will measure the results of the government’s voluntary pre-release process. The performance figures, rankings and projected resource savings cited here are Google’s claims; independent validation and fuller benchmark comparisons were not included in the supplied material.
Feedback Before Wider Release
Google says it will expand access gradually, collect feedback from early testers and continue work on safeguards. The company has said it intends to offer Argon to developers, enterprises and consumers as soon as possible, but has not announced a timetable. The next concrete milestones are expected to be updates on access, testing and any changes to the release safeguards or pricing.
Key Questions
What is Gemini 4 Argon?
Google describes Gemini 4 Argon as a frontier model designed for long, complex tasks in software engineering, enterprise knowledge work and defensive cybersecurity.
Who can use Argon now?
Google says the initial rollout is limited to trusted cyber defenders through its Fairwind Program, along with internal Google teams.
When will Argon be generally available?
Google has not provided a release date. It says wider access for developers, enterprises and consumers will follow early testing and further guardrail work.
What does Google say Argon can do for cybersecurity?
Google says Argon can autonomously find, validate and patch critical software vulnerabilities. Those capabilities and the company’s reported performance figures have not been independently established in the supplied announcement.
How much will Argon cost?
Google listed introductory pricing of $2 per million input tokens and $10 per million output tokens. Cached input tokens are priced 95% below the standard input rate.
Source: hn
Fall Picks
fall essentials
As an affiliate, we earn on qualifying purchases.
