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 initially rolling out to trusted cyber defenders through its Fairwind Program. The company reports strong results in coding, enterprise work and defensive cybersecurity, but those figures are company-reported and broad availability has not begun.
Google announced Gemini 4 Argon on Sept. 30, saying the model is beginning a limited rollout to trusted cyber defenders through its Fairwind Program. The company says Argon is designed for long, complex workflows in software engineering, enterprise tasks and cybersecurity defense; broader access will follow as Google gathers feedback and develops safeguards.
Google describes Argon as a model built to sustain reasoning across long-horizon tasks, including coding, legal and finance work, and cybersecurity defense. It says the model will initially be available to a selected group rather than to developers, enterprises or consumers generally. Google said it is participating in a U.S. government voluntary process for pre-release model access while it gradually expands testing.
The company says Argon has a maximum output limit of 1 million tokens, up from 64,000 in the previous model. Google also announced introductory API prices of $2 per million input tokens and $10 per million output tokens; cached input tokens are priced at a 95% discount from the input rate. The source does not give a date for general availability.
Google reported internal results in several areas. It says Argon helped researchers beat a published baseline for a quantum computing subroutine by 40% in minutes, and that memory optimizations identified by agents could free more than 300 TiB once rolled out. These are company-reported examples, with some outcomes described as estimates or dependent on rollout. Google says migration work affecting critical systems is undergoing automated and manual audits, emulation testing and review.
A Limited Test for High-Stakes Work
The announcement pairs claims of stronger performance on complex knowledge and engineering tasks with a restricted first release. That approach matters because Google says Argon can autonomously find, validate and patch critical software vulnerabilities. Releasing such capabilities first to trusted defenders gives the company a way to test defensive uses while it continues work on safeguards and access controls.
For businesses and developers, the reported token limit and benchmark results suggest the model is intended for tasks that involve large amounts of material or many steps. But benchmark scores and internal case studies do not establish how Argon will perform across independent customer deployments. Its usefulness, costs and reliability outside Google’s reported examples remain to be seen.
Google’s Internal Argon Results
Google says Argon is already being used by thousands of its employees for specialized coding, research and writing. One example involves libgav1, Google’s open-source video decoder: the company says Argon agents replaced 32,000 lines of SIMD code in an existing Rust port with safe Rust, producing a decoder that ran 2.7 times faster than that port while preserving identical video output. Google says the result brings performance closer to the optimized C++ implementation.
The company also cites scores on external benchmarks: 77.9% on DeepSWE v1.1 for software engineering tasks, 51.3% on Zapier’s AutomationBench, and 91.7% on LVBench for long-video understanding. Google says Argon leads the Vals Index and reports strong performance on finance and legal evaluations. These figures describe specific tests, not a general measure of performance in every real-world setting.
““Safely releasing frontier capabilities at this level requires a phased approach.””
Access and Results Still Unsettled
Google has not announced when Argon will become available to developers, businesses or consumers, or how quickly the Fairwind rollout will expand. The announcement also does not specify the full membership or selection criteria for the initial tester group.
The performance figures and operational savings cited in the announcement come from Google or benchmark results it selected. The company did not provide independent validation for its internal examples, and some projected savings depend on implementation. Details of Argon’s safeguards, including how they will change between trusted testing and any broader release, remain unspecified.
Feedback Before Wider Release
Google says it will expand access gradually, collect feedback from early testers and continue iterating on safeguards. The company has said it aims to make Argon available to developers, enterprises and consumers as soon as possible, but has provided no release schedule. Further details on access, safeguards and independent evaluations will clarify how the model’s reported capabilities translate into wider use.
Key Questions
What is Gemini 4 Argon?
Google describes Gemini 4 Argon as a frontier model for complex, multi-step work in software engineering, enterprise knowledge tasks and cybersecurity defense.
Who can use Argon now?
Google says the initial rollout is for a set of trusted cyber defenders through its Fairwind Program. General access has not begun.
When will Argon be generally available?
Google has not given a date. It says it will expand access gradually while gathering tester feedback and working on safeguards.
What does Argon cost?
Google announced introductory API prices of $2 per million input tokens and $10 per million output tokens. Cached input tokens cost 95% less than the input rate.
Are the performance claims independently verified?
The announcement includes company-reported internal examples and scores on named benchmarks. It does not provide independent validation for the internal results, and benchmark scores apply to the specific tests described.
Source: hn
Fall Picks
fall essentials
As an affiliate, we earn on qualifying purchases.
