[Digital Today reporter Hwang Chi-gyu] TechCrunch reported on the 30th (local time) that OpenAI will initially roll out its cybersecurity-focused model, GPT-5.5 Cyber, to a limited group of users.
OpenAI CEO Sam Altman wrote on X (formerly Twitter) that the company will phase access to GPT-5.5 Cyber to \"key cybersecurity leaders\" over the next few days.
OpenAI has set up an online application process requiring applicants to submit their qualifications and describe intended use cases.
GPT-5.5 Cyber can perform penetration testing, vulnerability discovery and exploitation, and malware reverse engineering. OpenAI positions it as a toolkit for organizations to uncover security flaws and validate defenses, but cybersecurity officials warn those same capabilities could be misused by malicious actors.
Altman previously criticized Anthropic's selective release of Mythos as \"fear marketing.\" Some critics had called Anthropic's approach overblown, and TechCrunch noted that OpenAI appears to be following a similar selective-release strategy.
OpenAI said it will coordinate with the U.S. government and vet additional users with bona fide cybersecurity credentials to expand access to GPT-5.5 Cyber.
