I tested the paired adapters at three locations in my home using this procedure, running separate up and downlink tests at each location. I also ran simultaneous up/downlink and four-stream tests at Location A. All results have been entered into the Powerline Performance Charts database.
The chart below shows average uplink throughput for all three test locations for all 500 Mbps powerline adapters I've tested. There isn't a simple 200/500 filter in the Powerline Charts, so I created the chart using the Powerline Product Finder. I just set the Features > Maximum Link Rate filter to 500, then ticked the Uplink Throughput Compare Performance checkbox for all the products shown, then hit the [Compare Performance] link.
Average uplink throughput
And here is the downlink chart.
Average downlink throughput
The XAV5101 really outdistances the pack on uplink, but not for downlink. A look at the throughput vs. location plot below shows that this is mainly due to the 216 Mbps throughput measured in the Location A (same outlet) test.
Same-outlet tests sometimes risk getting lower than normal results due to signal overload. But the XAV5101 appears to not have this problem.
Uplink throughput vs. location
For downlink, the XAV5101's throughput is only about 8% higher than the next-best adapters in Location A and has about the same throughput as other products at the other two test locations.
Downlink throughput vs. location
The Simultaneous up/down test for a single pair yielded 206 Mbps of aggregate throughput as shown in the IxChariot plot below. You can clearly see the uplink vs. downlink preference. I ran a baseline check between the two computers (connected via Gigabit switch) used for this test to make sure that this was not something Windows was doing. The test also showed an up vs. down preference—800 Mbps vs. 693 Mbps—not as pronounced as that shown below
Simultaneous up/downlink test - Location A - IxChariot
Adding a second pair didn't find much more throughput, but it sure showed unequal throughput distribution among test streams! I again ran a test with the two computers connected via Gigabit switch and found similar behavior. I upgraded to the latest drivers and tried changing various settings in the Intel Ethernet adapter properties, but could not eliminate this behavior. I did find, however, if I dialed back the test script send rate so that total throughput did not exceed 1 Gbps that throughput divided equally and was rock-steady.
Simultaneous up/downlink test 2X - Location A - IxChariot
Use the Powerline Charts to explore further.
Since the larger XAV5001 has been discontinued, the XAV5101 (or more accurately, the XAVB5101 kit) is it, if you want to buy your 500 Mbps HomePlug AV adapters from NETGEAR and don't want outlets on them. NETGEAR's site also shows an XAVB5201 that's "coming soon". But I can't tell what the difference might be from the XAVB5101.
At any rate, I wouldn't buy the XAVB5101 simply because of its high uplink Location A performance. After all, who is going to use two powerline adapters plugged into the same outlet? If you set that result aside, the XAVB5101 looks like it will behave like all the other "AV 500" adapters on the market.