An independent test of four artificial intelligence models verified 10 previously unpublished mathematical problems: the ETH Zurich model solved six problems, while other publicly available systems ...