All things being equal, I imagine if the best single stage A/C unit (~16 SEER) were pitted against the best variable speed unit (~26 SEER) and both were set to full blast all the time, they would roughly match efficiency. Variable speed units gain efficiency over single stage units because they don’t have to start/stop as much. Granted, in high temperatures there is more time spent “on” and less start/stop. But even in Arizona heat, A/C doesn’t run full blast all the time. They may take a hit, but variable speed units will still gain a lot of efficiency over single stage units in the heat and may even respond to sub-degree variations.I did a follow up question with my installer, and I agree I should use EER. But I still disagree with your statement. The reason high efficiency units drop out rapidly at higher temperatures (Arizona especially) is because the system cannot maintain an easy 1 deg F change. So it has to use the larger compressor or both compressors, which brings them back to EER 13'ish. (Likely along with my unit as well but nowhere "neer" as much EER.) If you want to compare specs, be my guest. I'm sure this has a gradient effect and may be just a curve function of 2nd stage compressor duty cycle.
The XC20 units we have can run from 35-100% in 1% increments, and I haven’t seen them max out for long. In our hot Arizona September they used ~30% less power than the 13 SEER units we replaced.