So please don't scoff at the lowly Transmille.
I haven't made any adjustments just curious as to if I needed to get something that is better than nothing.
Not scoffing, just trying to point out a way to accomplish what you are trying to do. I do have opinions on the matter, but since I'm not a recognized expert on either metrology or calibration lab procedure, I'll keep them to myself. Oh wait, this is EEVBlog. Well, I'll tell ya......
You may know all this, but for anyone following along, the HF correction constants are done differently than the gain constants. I haven't worked on a K2000, but I have worked on a number of same-class meters and they all do the HF correction in a similar way. The normal AC measurement simply has the signal pass through the ranging, voltage dividing and gain stages, then to the TRMS converter and then the ADC. After the ADC spits out a number, the meter's MCU multiplies it by the calibration gain constant for that range. The HF correction constant is a frequency response adjustment in one of the gain stages--usually voltage controlled--and as each range is selected, the MCU commands a DAC to output a voltage to the circuit to obtain the desired response. The idea is to get things as close as possible at one point (50kHz) and hope for the best as the frequency goes up--and you can see by the specs that things rapidly get worse out to the 300kHz advertised response. Typically the meters are fairly accurate to 20kHz already, so setting the HF gain constant there would not be helpful. The expectation is that the meter, once calibrated with good HF correction constants, will extend that accuracy out to 50kHz plus, after which it gets worse and worse.
An AWG or similar in parallel with a more accurate DMM would be how I would do it, but I don't have to certify anything. I don't know what procedures would be acceptable under ISO 17025. I have a calibrator, much older but similar in capability to yours--but its HF AC accuracy is not much better than a Siglent AWG in practice. The AWG allows me to verify the performance out to 300kHz as well.