Gemini 4 Argon
Google announced Gemini 4 Argon for long-horizon coding, knowledge work and cyber defense, with initial access limited to trusted defenders.
What happened lately
Gemini 4 Argon timeline
A restricted frontier-model rollout
Google announced Gemini 4 Argon on September 30, positioning it for long-running software engineering, professional knowledge work and cybersecurity defense. The company says initial access is going to trusted cyber defenders through its Fairwind Program. Developer, enterprise and consumer access is planned later, but Google has not announced a general release date. The model is therefore not broadly available to API customers or Gemini subscribers today.
Capabilities and safeguards
Google says Argon can produce up to one million output tokens in a single trajectory, compared with a prior 64,000-token limit. It reports strong results on coding and enterprise benchmarks and describes internal uses for code migration and vulnerability remediation. Those are company-reported evaluations and examples, not independent proof that the model will work equally well in other environments. Google says it is testing safeguards for cyber misuse, prompt injection and unintended agent actions before a wider rollout, while giving selected defenders access without its usual cyber guardrails.