Google limits Gemini 4 Argon rollout to cyber partners
The company says its new model can identify and patch software vulnerabilities. Initial access is restricted to selected Fairwind Program members.
Google has announced Gemini 4 Argon, a model it says is built for complex reasoning across finance, software engineering, coding, creative writing and cybersecurity defence. The company is beginning with selected members of its Fairwind Program, which serves governments and trusted partners seeking access to Google’s advanced cybersecurity models.
Google says Argon can autonomously find, validate and patch critical software vulnerabilities. Engadget reported an early demonstration in which the model identified a vulnerability in healthcare software used by hospitals, but the report provided no details on validation or disclosure.
The initial release is limited. Google plans to extend access to developers, enterprises and general users, with paid API customers and Google AI Ultra subscribers expected among later users. The company has not given a date for broader availability.
Engadget also reported that benchmarking firm Artificial Analysis found Argon matched GPT-6 Astra on its Intelligence Index at a lower cost per task at introductory prices. Those comparisons and the reported benchmark results are attributed claims, not independently established here.
Google says it has used Argon in quantum computing research, codebase migrations and data-centre memory optimisation. The company also says it has added safeguards against prompt injection and unauthorised action as it rolls out the model.
