OpenAI to give Japan cybersecurity model

OpenAI is preparing to provide the Japanese government and some Japanese companies with an advanced cybersecurity AI model, according to reports published on May 22 by Japan Today and Asahi Shimbun. The reports said the U.S. company behind ChatGPT would soon make the model available in Japan for government use and for selected private-sector users. Neither report listed contract dates or named the companies involved. OpenAI has separately been expanding a restricted-access cyber program in 2026. The company said on April 14 that it was scaling its Trusted Access for Cyber program and introducing GPT-5.4-Cyber for verified defenders, then said on May 7 that GPT-5.5 and GPT-5.5-Cyber were being delivered to developers and security teams through that framework. ### What exactly is Japan getting? Asahi Shimbun said OpenAI plans to provide its “latest cybersecurity AI model” to some Japanese companies, while Japan Today said the arrangement also includes the Japanese government. (asahi.com) The published reports described the tool as an advanced model with stronger cybersecurity capabilities, but did not disclose pricing, deployment terms or a timeline for broader rollout. OpenAI has not published a Japan-specific announcement on its own site that matches those reports. (openai.com) The company’s public materials instead describe Daybreak as its cybersecurity offering for software defense and say organizations can contact sales or request a vulnerability scan. ### Is this the same thing as Daybreak? TechWire Asia reported on May 22 that OpenAI’s Daybreak cybersecurity offering uses GPT-5.5 models and Codex Security in workflows including code review, threat modeling and patching. (asahi.com) OpenAI’s Daybreak site describes the product in similar terms, saying it is designed to identify threats, generate patches and verify remediation across code and systems. (openai.com) OpenAI’s own wording stops short of naming a Japan deployment. But the overlap in function — software defense, vulnerability identification, patch generation and remediation checks — indicates the Japan reports are referring to the same cyber product family or a closely related deployment, based on the public descriptions. ### What can the model do inside a security workflow? OpenAI said Daybreak is intended to help teams “see risk earlier” and make software “resilient by design.” The company’s materials say the system can prioritize high-impact issues, generate and test patches inside repositories, and return audit-ready evidence to customer systems for remediation tracking. (techwireasia.com) The Hacker News, citing OpenAI’s product materials, reported that Daybreak uses Codex Security to build an editable threat model for a code repository, identify and test vulnerabilities in an isolated environment, and propose fixes. (techwireasia.com) CNBC separately reported that GPT-5.5-Cyber is trained to be more permissive on security-related tasks and is being rolled out in limited preview to vetted cybersecurity teams. (openai.com) ### Why is access being limited? OpenAI said in February that Trusted Access for Cyber was designed as an identity- and trust-based framework to place enhanced cyber capabilities with approved defenders while reducing misuse risk. On April 14, the company said it was preparing for “increasingly more capable models” over the following months and was fine-tuning models specifically for defensive cybersecurity use cases. Japan Today also reported last month that OpenAI would release its latest cybersecurity model only to a limited number of partners, after Anthropic similarly restricted release of a cyber system. (thehackernews.com) That report said the restricted approach reflected concern about AI-enabled cyber misuse. ### What remains undisclosed? May 22 is the only date attached publicly to the Japan reports, and the named Japanese corporate recipients have not been identified in the available coverage. (openai.com) OpenAI’s Daybreak site currently points prospective users to a sales contact and a vulnerability-scan request form, while the company’s security news page continues to list updates on GPT-5.5-Cyber and related cyber-defense programs. (openai.com) (japantoday.com)