Sam Altman has a new cybersecurity tool called Cyber, and he’s decided that only the right people should get to play with it. This is the same Sam Altman who, not long ago, publicly mocked Anthropic for doing exactly that with its own tool, Mythos.
On Thursday, Altman announced on X that OpenAI would begin rolling out GPT-5.5 Cyber “to critical cyber defenders” in the coming days. Aspiring defenders can apply via a form on OpenAI’s website, where they must submit their credentials and explain how they plan to use the tool. The application suggests Cyber can handle penetration testing, vulnerability identification and exploitation, and malware reverse engineering - basically, a Swiss Army knife for finding security holes and testing defenses. The catch? The same kit could be used by the bad guys, which is why OpenAI is being careful.
When Anthropic restricted access to Mythos, Altman called it “fear-based marketing.” Critics agreed, saying Anthropic’s rhetoric was overblown. But irony, as ever, is a dish best served cold: an unauthorized group reportedly got their hands on Mythos anyway. OpenAI says it’s working to make Cyber more widely available by consulting with the U.S. government and identifying users with legitimate cybersecurity credentials. Because if there’s one thing that screams “not fear-based,” it’s government consultation and an application process.