Google launches Gemini 4 Argon for cyber defense and complex work

Google introduced Gemini 4 Argon, its next frontier model for long-horizon software engineering, professional knowledge work and cybersecurity defense.

The rollout is deliberately phased. Google says Argon is going first to trusted cyber defenders through its Fairwind Program while the company gathers feedback, works through voluntary U.S. government pre-release access, and strengthens guardrails before broader availability for developers, enterprises and consumers.

The practical shift is depth of work. Google says Argon expands the output limit to 1M tokens, giving the model room to carry out much longer reasoning and generation trajectories. Internally, Google says teams have used it for specialized coding, large codebase migrations, memory optimization across data centers and research workflows.

Cybersecurity is the sharpest initial use case. Google says Argon can help defenders find, validate and patch critical vulnerabilities, and that Wiz is already using it through Scan for Good to protect public infrastructure. Google also says the model ties for first on CWE-bench v1 and improves on earlier Gemini cyber models in vulnerability discovery.

For builders, the launch matters because Google is positioning Gemini 4 not as a chat upgrade alone, but as a model for sustained agentic work. Introductory API pricing is 2 USD per million input tokens and 10 USD per million output tokens, with cached input tokens discounted.