OpenAI Gives South Korea Access to Specialized Cyber AI Model

OpenAI has launched a "Korea Cyber Action Plan" to bolster the nation's cybersecurity capabilities. The plan includes providing the Korean government with access to "GPT 5.5-Cyber," a specialized AI model for security. This makes South Korea the third country, after the United States and Canada, to receive such access, and the first in Asia, alongside Japan.

Jason Kwon, OpenAI's Chief Strategy Officer, stated during a press conference in Seoul that the company aims to be a trusted partner for South Korea. He emphasized OpenAI's goal of providing advanced technology to reliable institutions and promoting responsible AI use. Kwon highlighted South Korea's strong AI utilization capabilities, citing its semiconductor industry, talent pool, infrastructure, and public sector.

“

OpenAI wants to be a trusted partner for Korea. Our goal is to give more technology to more trusted institutions and to use AI responsibly.

OpenAI has been actively seeking collaborations in South Korea since signing a memorandum of understanding with the Ministry of Science and ICT in October. Recent engagements include a workshop on AI security with various government agencies and a high-level meeting that formalized South Korea's participation in OpenAI's Government and Institutions Trust-based Access Program (GTAC). Through GTAC, participating Korean entities will gain access to cutting-edge AI models like GPT 5.5-Cyber.

“

Unlike Anthropic, we operate the program broadly because we want as many people as possible to be able to defend themselves as quickly as possible.

Kwon contrasted OpenAI's open approach with that of Anthropic, which he described as providing access to its "Claude Mitos" model through a more closed "Project Glasswing." He explained that OpenAI's broader program aims to empower more individuals to defend themselves quickly, given the accelerating pace of cyber threats. The initiative seeks to enable trusted defenders to build capabilities, identify vulnerabilities using AI, and deploy patches to prevent the spread of cyberattacks.

In addition to government access, OpenAI is in discussions with South Korean private companies about participating in the TAC program. This move positions South Korea alongside the U.S. and Canada as a key recipient of advanced AI security technology, with ongoing discussions also underway with EU nations. The development reflects a growing global race among tech giants to develop AI models specifically for cybersecurity.

“

Given the speed of AI development, cyber threats will only grow. We want trusted defenders to quickly build defense capabilities, find vulnerabilities with AI, and patch them to stop cyberattacks from spreading.