OpenAI Is Giving Ukraine Free Access to Daybreak for Civilian Cyber Defense

OpenAI is working with Ukraine’s Ministry of Digital Transformation to provide Ukrainian teams with Daybreak tools for defending civilian infrastructure. The company says the tools can help defenders identify software vulnerabilities and develop and test fixes faster; the BBC reports that access is free and includes GPT-5.6 Sol. The announcement describes a planned defensive capability, not a measured reduction in cyberattacks. 421

What Ukrainian defenders would use Daybreak for

Daybreak gives authorized cyber defenders access to AI for reviewing older software, investigating suspicious activity, validating vulnerabilities and testing repairs. Those tasks matter for civilian services such as hospitals and power systems, which are among the infrastructure the partnership aims to protect. OpenAI has also announced training and technical support for Ukrainian critical-infrastructure teams. 42134

GPT-5.6 Sol is one of the models available through Daybreak Blue, OpenAI’s access tier for authorized defensive work. The available announcements do not specify Ukraine’s access tier or say that it will receive Daybreak Red’s specialized GPT-5.6-Cyber model. Nor do they establish that CERT-UA, Ukraine’s national incident-response team, is a named recipient, even though the BBC cites CERT-UA’s count of nearly 6,000 cyber incidents in 2025. 10421

How this builds on OpenAI’s European work

OpenAI says it has already provided cyber models to defenders in France, Germany and Poland. It cites ENISA’s use of the models to find vulnerabilities in software used by EU institutions that were subsequently fixed. OpenAI also says CERT Polska used its models to help find six flaws in third-party router software, after which the vendor issued fixes. Those examples show the kind of work Daybreak could support in Ukraine; they are not evidence of outcomes there. 4248

The wider AI-security stakes

The partnership comes amid concern that AI tools can aid attackers as well as defenders. Anthropic has reported disrupting alleged misuse of Claude by Russia-linked developers working on autonomous attack-drone software. That account should not be read as proof that the resulting software was deployed. 1830

Private technology firms have already played significant roles in Ukraine’s wartime infrastructure and defense, according to reporting and analysis. Working with Ukrainian defenders could also give a provider operational experience, but that possibility does not establish OpenAI’s motive. The immediate test for Daybreak is whether Ukrainian teams can use it to find and fix consequential weaknesses—not simply whether access has been announced. 223242