OpenAI has acquired access to the FrontierMath: Open Problems validator, a tool developed by Epoch AI for verifying solutions to unsolved advanced mathematical problems. This validator, which includes dedicated computer programs for each problem, allows OpenAI to check the validity of solutions generated by its AI models. The validator is available for purchase by any organization, with costs primarily covering the labor-intensive development by mathematicians. Epoch AI stipulates that any valid solutions discovered must be reported, granting joint authorship rights to the problem's original author and the discoverer. OpenAI previously funded the creation of the FrontierMath benchmark, but the Open Problems component was independently developed by Epoch AI with initial funding from Schmidt Sciences.