Quick answer
Google announced Gemini 4 Argon on September 30, 2026, with initial access for trusted cyber defenders and testers. The company says wider availability will begin with paid API customers and Google AI Ultra subscribers, but has not given a public release date. Google describes a one-million-token output limit; that is not the same as a one-million-token input context window. The exact model is not listed in Chat AI's current public model directory, so availability in Chat AI is unverified.
Quick answer: Gemini 4 Argon is announced, but access is still limited
Google announced Gemini 4 Argon on September 30 as a model for long-running work in software engineering, business knowledge tasks, and cybersecurity defense. Its first rollout is to a limited group of trusted cyber defenders and testers. Google says later access will start with paid API customers and Google AI Ultra subscribers, followed by broader developer, enterprise, and consumer access. The announcement does not provide a date for those later stages.
Sources: Google, Google DeepMind
The one-million-token figure describes output, not input context
Google says Argon's maximum output token limit is increasing from 64,000 to one million tokens to support longer task trajectories. This is an output limit; the announcement does not state that the model accepts a one-million-token input context. The distinction matters when comparing model specifications or estimating how much source material can be provided in one request.
Sources: Google
Cybersecurity access is being staged through a trusted program
Google says Argon is first rolling out through Fairwind to trusted cyber defenders. Its earlier Fairwind description explains that the program serves selected government, critical-infrastructure, and technology partners, with operational rules for participating organizations. Google's Argon announcement also says it is continuing safeguard work before wider access, including misuse defenses, prompt-injection testing, monitoring for actions outside a user's intent, and more isolated environments for high-risk evaluations.
Google's benchmark results are provider-reported, not an independent app comparison
Google publishes benchmark scores and internal workflow examples alongside the announcement. They are useful for understanding what the company chose to measure, but they are not an independent evaluation or a guarantee of performance on a particular task. A practical comparison should wait until access is available and test the same representative work, with the same tools and review criteria, across the products being considered.
Sources: Google, Google DeepMind
Gemini 4 Argon is not verified in Chat AI
The exact Gemini 4 Argon name is not in Chat AI's current public model directory, so this article does not claim that the model is available in the app. Check the directory or model selector for current listings. For choosing among apps, compare exact model access and test the same task rather than relying on a provider announcement alone.
Frequently asked questions
What readers usually ask
Can the public use Gemini 4 Argon now?
Google says the initial rollout is to trusted cyber defenders and testers. It plans to start broader access with paid API customers and Google AI Ultra subscribers, but has not announced a public date.
Does Gemini 4 Argon have a one-million-token context window?
Google's announcement describes a one-million-token maximum output limit, not a one-million-token input context window.
What is Google's Fairwind program?
Google describes Fairwind as a limited-access cyber-defense program for trusted government, enterprise, and cybersecurity partners. Argon is beginning its rollout through a cohort of cyber defenders and testers.
Is Gemini 4 Argon available in Chat AI?
Its availability is unverified. The exact model name is not listed in Chat AI's current public model directory, so check the live catalog rather than assuming access.
Evidence