Google launches Gemini 4 Argon for cybersecurity and complex workflows
Alphabet parent company Google has launched Gemini 4 Argon, a new AI model designed for defensive cyber work and long-horizon tasks.
Capabilities and token limits
The model features an output limit of one million tokens, a sharp increase from previous versions. Google states this capacity allows the system to think more deeply on difficult problems. Staff members have already used the tool for daily debugging and codebase migrations.
Performance benchmark results
Argon scored 77.9% on the DeepSWE coding benchmark, placing it ahead of other major models. It also achieved a score of 51.3% on Zapier's AutomationBench for business function execution. The system demonstrated state-of-the-art performance of 91.7% on tests measuring long video understanding.
Distribution and access
Google is rolling out the model to a select group of cyber partners through its Fairwind Program. The company claims the system can autonomously find, validate, and patch critical software vulnerabilities. Access is currently restricted to trusted defenders and specific Google teams without general guardrails.
2 outlets, different leads
-
TechCrunch
Google releases Gemini 4 Argon, called its most powerful model yet
-
The News International
Google launches Gemini 4 Argon as its new frontier model
Reported by
2 independent outlets. Headlines as published. Links open the original report.