If I'm calculating this correctly, the theoretical best performance for my particular ZWO AM5 mount with a maximum period error of 23.6" is 0.136 arc-second error for a 1 second guide duration. So, 0.136" error is what's possible with perfect seeing, no atmosphere, in Earth's orbit somewhere. Here on the ground, ZWO has specified that all AM5 mounts should perform with 0.5" to 0.8" total error or better, given average seeing and atmospheric disturbance. And that seems to be exactly what I am experiencing.
For a detailed walk-through of the periodic error report values and calculations see 10 things you need to know about the custom AM5’s PE Test Report provided by ZWO