What happened
On September 30, 2026, Koray Kavukcuoglu, SVP at Google DeepMind and Chief AI Architect, announced Gemini 4 Argon, Google's new frontier model. According to the post, it is built to sustain deep reasoning across long, complex workflows, with frontier performance in real-world software engineering, enterprise knowledge work (legal and finance) and cybersecurity defense.
A phased launch
The unusual part: Argon was not released to the public right away. It started with a set of trusted cyber defenders through the Fairwind Program. Google says it will use early testers' feedback to strengthen guardrails and then expand access, starting with paid API customers and Google AI Ultra subscribers.
Announced pricing
- Introductory price: $2 per million input tokens and $10 per million output tokens;
- cached input tokens at 95% off;
- after the introductory period the price rises (Google cites $4 per million input).
Why it matters
It is another sign that models with strong offensive capability are going to defenders first, a pattern seen at other labs this year. For businesses, the window between "defenders have the tool" and "everyone has it" is the time to fix what is already known to be vulnerable.
What to do now
- If your security team works on critical infrastructure, follow the Fairwind criteria.
- Budget for the post-introductory price, not just the promotional one.
- Use the restricted-access period to close known vulnerabilities before general release.