I don't know one way or another but there is a credible allegation that the "AGI" was training on the (very extensive) test set that these two mathematicians produced.

If that is true then this seems to be, again, a case of AI producing an interpolation over data it has seen before. Everything about openAI's behavior indicates that they were using the transcripts as input. Why not have the AGI choose a different Millenium prize problem?