The White House's voluntary AI cybersecurity testing framework reportedly excludes open models entirely, and explicitly can't be used to evaluate them.
Continue to AI University →