Following Sam Altman’s criticism of Anthropic for controlling access to its cybersecurity tool Mythos by only providing it to chosen users, he disclosed that OpenAI will adopt a similar approach with its rival tool, Cyber.

In a post on X on Thursday, Altman announced that OpenAI plans to start deploying GPT-5.5 Cyber “to essential cyber defenders” in the coming days. OpenAI has a form on its website where individuals can submit details about their credentials and intended usage to obtain access.

The application suggests that Cyber can execute tasks such as penetration testing, identifying (and exploiting) vulnerabilities, and reverse engineering malware. It is designed to serve as a toolkit to aid companies in identifying security weaknesses and testing defenses. There are concerns that the toolkit could be misappropriated by malicious actors.

When Anthropic similarly limited access to Mythos, Altman referred to the strategy as fear-based marketing. Some detractors echoed this sentiment, claiming that Anthropic’s messaging was exaggerated. Ironically, a group unauthorizedly reportedly gained access to Mythos nonetheless.

OpenAI asserts that it is striving to broaden access to Cyber by consulting with the U.S. government and identifying additional users with valid cybersecurity credentials.