The arrangement gives Kyiv access to GPT 5.6 Sol in a version tailored for cyber defense, The Insider reported. The model has been publicly available since July, but Daybreak runs it with relaxed safeguards so that specialists can search for vulnerabilities and fend off attacks.
The system is designed to locate flaws in digital systems quickly and to help develop fixes. Ukraine's national computer emergency response team, CERT-UA, recorded almost 6,000 cyberattacks against the country in 2025.
OpenAI announced the deal on September 23 on the sidelines of the UN General Assembly in New York. George Osborne, the former British chancellor of the exchequer who now heads OpenAI for Countries, said protecting civilian infrastructure means defending it against both physical and digital attacks so that people can continue to live, work and use vital services.
Jamie MacColl, a senior research fellow at the Royal United Services Institute (RUSI), noted that Ukraine already has AI tools from OpenAI's competitors, Google among them. Western technology companies support Ukraine partly out of altruism and partly because they obtain valuable data and intelligence during an active conflict, he said.
OpenAI's rival Anthropic tightly restricts access to its models, saying they are too powerful and could fall into the wrong hands. OpenAI, by contrast, has allowed a number of European companies and British banks to use its most advanced cyber tools.
Anthropic said earlier that Russia-linked hackers had used its Claude AI agents to attack Ukraine. They downloaded the contents of mailboxes at at least two manufacturers of drone components, attacked a military drone maker and stole a software development kit for its computer vision system.
In May it emerged that Ukraine is using AI software from Palantir Technologies, a major Pentagon supplier. Then defense minister Mykhailo Fedorov said the cooperation had produced a system for detailed analysis of air attacks and introduced tools for handling large volumes of intelligence data.