You know, there isn't exactly a clearinghouse for weapon performance stats. Every time someone starts talking about the crummy M16, someone else reminds them that it was the most reliable in recent NATO tests. Then you bring up Vietnam and Afghanistan and the debate rolls on.
As for UZI/MP5 debates, I don't think either of them are as good in all sorts of crud as the Sterling, but I have no data to back that up.
Military rifle reliability is funny. The standard line about AK reliability is taken for granted, but I've heard of them jamming. And a friend in the Danish Army reported incredible levels of reliability with the fairly precise G3. So maybe loose tolerences don't tell the whole story.
I always take individual reports on weapons with a bit o' salt. I've never heard anything that would convince me that an AUG is any less reliable than any of the other 10 or more western gas driven .223s. If there is a downside, it is likely shared by all of those weapons since the critical components are so similar.
AUG, FNC, AR18, M16, Galil, Tavor, G36, AR70, 550, C7, MD-2, K2, AC556, Valmet, Stoner 63 all do the about the same stuff in the same way. They all are expected to have a certain level of accuracy, too. So unless the designer screws up (SA80), they all come out working about the same, their reliability dictated largely by the clearences necessary to acheive the desired accuracy. Almost all are VERY accurate.
Personally, I'm in the delayed blowback club. HK33, Cetme L, FAMAS. These guns are reliable and accurate, despite having received so much less development than the gas guns. Oh well.
