Sam Altman fired back at Anthropic after the company limited its Mythos cybersecurity tool to a select group, labeling the move a scare tactic. In response, OpenAI revealed it will now gate access to its own tool, Cyber, as part of a rollout for “critical cyber defenders.” The announcement appeared on X on Thursday.

Cyber promises to perform penetration testing, identify and exploit vulnerabilities, and reverse engineer malware, according to an application OpenAI hosts on its website. Users must submit credentials and a use‑case plan to qualify. The company warns that misuse by malicious actors could turn the kit into a weapon against its own defenses.

Altman’s critique of Anthropic’s Mythos gatekeeping sparked debate over whether defensive tools should be broadly available or tightly controlled. Critics argued the rhetoric was overstated, especially after an unauthorized group reportedly bypassed Mythos. OpenAI now claims it is consulting with U.S. government officials to identify legitimate security professionals before expanding Cyber’s reach.

The move signals a broader trend in AI‑driven security tools, where vendors balance commercial potential against the risk of weaponization. By tightening access, OpenAI aims to prevent the technology from falling into the wrong hands while still offering advanced capabilities to vetted defenders. The company’s next steps will reveal how the industry negotiates this delicate trade‑off.