I spent a lot of time with Crimson over the weekend. I A/Bed it against my usual listening. A12t, UERR, Helios SE and Annihilator 2023. Between all of them the peaks and peaks were very much dependant on the tips I used. When I use the velvet tips I found some harsh peaks at 9k and 16k. When I switched to Coreir brass their was a larger dip at around 7-9k that brought it much close to linear. The interesting thing was that when I measured them on my 711c, the dips and peaks I noticed were not on the graph. I think using the 711c as evidence of peaks and dips post 5khz (honestly maybe lower) is a fools errand. The peaks and dips move based on variables like insertion depth, bore size, nozzle angle tip material, ear anatomy etc. It goes on.
I don't think audio science has advanced enough for a $100 coupler for Aliexpress to accurately portray the really high range treble that you're looking at. We can get broad strokes, but we need to use our ears in combination to validate the data these tools give us.