I've done some comparisons with my beloved Nikon d500, x-h2, and x-t3 (the latter basically as a control, although it did quite well) using the Nikon 500mm pf (with the Fringer AF adapter on the Fuji cameras) and the Fujifilm 70-300. Overarching goal is to see if I should keep my new x-h2 and possibly sell my d500. On paper, the 40mp x-h2 sensor with the 500mm lens should be the best. The x-t3 with the 70-300 should be the worst, only because I can't put that lens on the d500 (which has only 20mp compared to the 26mp of the x-t3).
I am sure there are problems with my testing, and I welcome feedback along those lines, but my goal was just to do comparisons that mirrored my real-world shooting conditions.
I'm starting with jpeg comparisons in this thread because I am having issues with the x-h2 raw files (discussed in another thread). Will get back to that at some point.
The Fuji and Nikon jpeg files are both as standard as I can get them, "standard" picture control in Nikon, "provia" in Fujifilm, everything at default values. There certainly are differences in how Nikon and Fuji engineers have chosen to pick defaults, but, what are you gonna do. I think the results are clear enough with the jpegs that the comparisons are still useful, to me at least.
The light was identical (within reason) in all comparisons. I did this before dawn with two LED spotlights on the dollar bill (these lights were about 5 feet away) and one overhead light, on my bathroom wall if you must know.
The five combinations being compared are:
1. x-h2 with 500mm pf.
2. x-h2 with 70-300mm at 300mm.
3. d500 with 500mm pf.
4. x-t3 with 500mm pf.
5. x-t3 with 70-300mm at 300mm.
They are always presented in that order, left to right.
All tests wide open at f/5.6.
Spot metering on George's face.
Handheld with VR/IBIS/OIS on.
AF-C on George's face.
The target is 14' away from me, at camera height.
Continuous shutter.
I took 3 bursts in each scenario, refocusing between bursts, and picked the best picture. There were usually about 10-12 photos to choose from.
My overall goal was to replicate my real-world shooting scenarios.
The height of George's likeness on a dollar bill is smaller than the smallest bird I've ever shot, save maybe hummingbirds: less than 1.5". So, we are talking about taking pictures of tiny birds at 14', quite optimistic.
The dollar bill itself is a bit over 2.5" tall.
I shot at 1/30, 1/60, 1/125, and 1/250, going from optimistic to reliable for birds on sticks.
I scaled the pics so that the subject (George's face standing in for a bird) was the same. I always started with the x-h2 and 100% and sized the others to match. Once I got the sizing acceptable, I used the same percentages for all comparisons (from left to right, 100%, 180%, 140%, 123%, 226%).
ISOs varied a bit given differences between systems, cameras, and what exactly was included in the "spot" that I was trying to land on George's eyes/nose. The x-t3 tended to have the largest ISOs but they were generally all within a stop for a given shutter speed.
First, at 1/30 (base ISOs on all combinations, and these pics are a touch overexposed (maybe 1/3 of a stop), really testing IBIS, VR, etc. here:

Seems to me that the x-h2 + 500mm pf is the clear winner (first shot), and the two Fuji cameras with the 70-300 are about the same and clear losers (second and fifth shots). The x-t3 + 500mm pf (fourth) and the d500 (third) with that same lens are quite close. I'd give the slight edge to the x-t3 (fourth) as I think we would expect given 26 v 20 mp. Really impressed with the sharpness of all pictures frankly, at 1/30! (although, keep in mind that in some cases there might have been only one or two good shots out of a dozen).
Now 1/60:

Similar, but the d500 + 500mm (third) falls noticeably behind the x-t3 + 500mm, presumably because the VR couldn't get me good shot out of a dozen. Maybe the sharp shot for the d500 + 500mm lens at 1/30 was a bit of a fluke? x-h2 (first) beats the x-t3 (fourth) with the 500mm lens but...it's darned close. Looks smaller than the difference at 1/30, maybe the IBIS is helping the x-h2 (first shot) at the longer shutter speed of 1/30.
1/125:

Dang that x-h2 + 500mm shot (first) is sharp, probably the sharpest shot in all the comparisons. X-t2 (fourth) not far behind and noticeably better than the d500 (third) when both have the 500mm lens attached. Both 70-300 shots (second and fifth) are worst, but the d500 (third) did not do well here. This is the danger zone for shutter speed for the VR problem of this lens that you either believe in or don't. I do.
[as an aside, if you are curious, here's a shot where I doubled the size of each picture for the 1/125 comparison; when I look at the detail in George's left eye in the x-h2 + 500mm image, at 14', I am thoroughly impressed]

And 1/250, all combinations should be solid in terms of IBIS/VR at this shutter speed.

Not much at all between the x-h2 (first) and x-t3 (fourth) shots with the 500mm lens here. d500 (third) has come back into shape and is again clearly better than the two shots with the 70-300 (the second and fifth).
I will say that the differences between the best and worst pictures was obvious at "fill" size, "fit" size, and 50% in LR. To tease apart the x-h2 and x-t3 pictures took 100% views or more.
Anyway, there you have it. Love to hear what you all think.