The LM4562 and the AD826 is a voltage Feedback OP Amps. They are pin and voltage compatible but may not work as designed based on the circuit around on the Op Amp. The OPA2134 is a fet based Op Amp.
I didn't say I heard an improvement on either of the LM chips, As a matter of fact, I thought the LM4562 sounded pretty much the same as the OPA2134.
The LM6172 had different sound to it but dont necessarily know that its any better or worse than the OPA2134 as the only tool I was using to listen to this was my ears, a pair of matched SDC mics, one on the unmodified channel and one on the modified channel and listened to some acoustic guitar and swapped the channels back and forth . All the statement was intended to say was I plugged in the chip, it passed audio and it sounded different than the stock chip. Not better or worse. Ive not made those determination yet.
The circuit may just be forgiving enough use some of these without modification but till I see it on a scope, I cant give you a definitive answer.
Edit: First clip.
http://thesupremeruler.com/downloads/mpagoldtest.mp3
The Setup: Matched pair of Rode NT5s both going straight into the ART MPA Gold, straight into the DA converter. No EQ, compression or effects. One channel of the ART has been modified with the LM6172 but I wont tell you which till later. The Tubes are stock. We'll change the tubes and record an AB from stock tube to NOS Mullard in the coming days.
As for the LM6172, it played fine on acoustic but didnt work so well on Bass DI. The gain was very erratic. The conclusion: This chip wont work properly without circuit modification. Could be as easy as some decoupling caps but till I get more in depth on the circuit, I wont know.