After a bit more tinkering, it seems that this slight magenta cast in the 20-25% gray range was somehow being introduced by setting the gray balance correction to "Iterative" (instead of just "On" as is the default.) I calibrated Monitor 1 with Iterative, and Monitor 2 without, then only Monitor 1 demonstrated the problem. After recalibrating Monitor 2 with Iterative, it demonstrated exactly the same issue. (But I have to say, they did both match exactly... bizarre magenta problem and all.)
So I reverted the gray balance thing to default ("On"), enabled the monitor RGB control option in the calibration software, then recalibrated both of the monitors, carefully tinkering with the monitors' RGB settings until the xy values, brightness, and color temp read by the colorimeter were right on target.
This seems to have given me reasonable results. Any slight differences I notice seem to change with viewing angle, and at this point I'm pretty confident that these displays are calibrated about as well as they possibly can be. I'm also pretty sure that there's nothing defective about either monitor and that any issues (like 0-13% gray all appearing as black, viewing angle distortion, etc) are just normal limitations of their technology/design.
I still plan to get a more suitable wide gamut monitor of some sort, but at least now I don't feel nearly as much urgency to do so....
So I guess the main lesson here is that "Iterative" isn't necessarily better than default even if it's theoretically more rigorous. (I tried to find some docs on exactly what the "Iterative" setting does and how it works, but I was unable to do so.) So now I'm just left wondering: Why?