Some of the GAP/R legend needs a bit of clarification:

"You should also know that the 12AX7 tubes used in these modules were hand selected for match and stability, and were pre-aged before shipment"

None of the GAP/R circuits give a rat's rump about "matching" & the tubes, 12AX7 & otherwise were never matched. They were selected for key / critical characteristics, but not matched for anything, not even date codes.

One of the things that they were selected for was µ – these being vacuum tubes, all specifications are "typicals", meaaning a typical value that you might never actually see in your lifetime  ± at least 20% for just about every number in whatever units of measurements & that only on a good day. What the nimble-fingered elvesses (elviatrixae ?) @ GAP/R cherry-picked were the top  µ end of the distribution. If you run up a few GAP/R 12AX7 on a curve tracer or calibrated transconductance tester, you will find (the good ones at least) all crammed up near the same high end of the  µ distribution. Don't expect the two triode sections of a 12AX7 to be  µ-matched very well, though;  ± 5% is about as close as you're gonna get & that is a natural consequence of selecting for the top tail of the  µ distribution bell curve, not a deliberate choice. Mayhaps a bit more simply, if both triode sections are already towards the top end tail of the bell curve distribution, then both halves will be naturally closer than "normal" to each other.

Another characteristic that the genuine GAP/R tubes were selected for was low grid current, essentially using the tube itself as a Bayer-Alperd ionization vacuum gauge. I don't know what their Go/Nogo limits were, but this not even specified as a typical characteristic will really run all over the place – by at least 4 orders of magnitude. The tubes (all of 'em, not just 12AX7s) were aged under power & bias for about a week before they were tested / selected for use, then stuffed in their new homes. That was run through its paces, glued if it passed, & shipped without any burn-in at all – they either passed-sealed-shipped or went back to rework for another run at the brass ring. Noise, microphonics, shape/style/number of getters, shape/style/number of micas, shape, size/color of plates, base pin finish / plating, date codes, & everything else that the golden-eared audiophiles are so enamored with were never tested or selected for.