Details
- Google announced Gemini 4 Argon, a new frontier model designed for deep reasoning across complex, long-horizon workflows.
- The model targets real-world software engineering, enterprise knowledge work in legal and finance, and defensive cybersecurity.
- Argon expands its output limit to 1 million tokens, up from 64,000, enabling longer code, document, and multi-step workflow generation.
- Initial access is limited to trusted cyber defenders through Google’s Fairwind Program, which supports testing and vulnerability remediation.
- Google plans broader availability for paid API customers and Google AI Ultra subscribers before expanding access to developers, enterprises, and consumers.
- Introductory pricing is set at $2 per million input tokens and $10 per million output tokens, with cached input tokens discounted by 95%.
- Google says Argon can autonomously identify, validate, and patch critical software vulnerabilities, subject to safety testing and staged deployment.
- The announcement places Gemini 4 in direct competition with advanced models from OpenAI and Anthropic on enterprise, coding, and cybersecurity workloads.
Impact
Argon intensifies competition among Google, OpenAI, and Anthropic by combining a substantially larger output window with enterprise and cybersecurity capabilities. Its introductory pricing could pressure rivals on high-volume reasoning workloads, while the restricted Fairwind rollout shows that advanced cyber functions are being commercialized alongside tighter safety and government review. The limited initial release may delay broad adoption, but staged access gives Google a way to gather operational feedback before exposing the model to larger developer and business audiences.
