OpenAI has announced that its cybersecurity tool, GPT-5.5 Cyber, will be rolled out selectively, echoing its previous criticism of Anthropic's approach to its own tool, Mythos. CEO Sam Altman previously criticized Anthropic for restricting access, labeling it as a form of fear-based marketing.
In a recent post on X, Altman confirmed that Cyber will be made available only to a limited group of users, specifically targeting critical cyber defenders. Interested parties can apply through OpenAI's website by providing their credentials and intended use for the tool.
Cyber is designed to assist organizations with various cybersecurity tasks, including:
- Penetration testing
- Vulnerability identification and exploitation
- Malware reverse engineering
This toolkit aims to help companies identify security vulnerabilities and test their defenses. However, there are concerns that such tools could be misused by malicious actors.
Despite his previous remarks about Anthropic's strategy, Altman stated that OpenAI is working to broaden access to Cyber. The company is collaborating with the U.S. government to identify additional users with legitimate cybersecurity credentials.
As the cybersecurity landscape evolves, the implications of such restrictions and the potential for misuse remain critical considerations for both developers and users of these tools.
