Did this say whether the reasoning models get this right more than the others? Was curious about that but missed it if it was mentioned.