Thanks for your clear reaction Ramses
My sample rate = 44.100 kHz  24bit
At 256 samples input latency = 6.44 ms output = 6.825 ms
At 128 samples input latency = 3.537 ms output = 3.923 ms
At 96 samples   input latency = 2.812 ms  output = 3.197 ms
An honest salesperson told me not to replace my old UCX with a new UFX because the only difference in my wallet would be tangible and therefore provide no tactile or audible improvement in latency

I wouldn´t exactly call him honest but maybe not very business-minded ;-).

Fact is, the latest UFX II / III sound much better than the old UCX, even though they don´t use the same converters as the actual ADI-2 Pro FS but the Adi-2 FS without the "pro".

If latency is your only interest, the improvement will be negligible, indeed.

Sound wise, you need to find this out yourself.
RME performs transparent conversion, so do not expect some "mojo jumps".
Maybe some subtle changes because of the different combination of DA converter and analog stage behind it, which brings the converter signal to (different) line levels. Some devices (the reference converter) offer to select different A/D and D/A filter. But also these are no night and day differences. Only subtle one. You should perhaps read KaiS findings about certain D/A Filter to better separate shakers from other instruments in the stereo panorama.
If it is something like this, what you are after, then you should perhaps check out the ADI-2 Pro FS R BE connected to your UCX's ADAT (optical SPDIF) port, to also get galvanic isolation.

The improvements of the UCX II over UCX you can read here, feature-wise:
https://forum.rme-audio.de/viewtopic.php?id=33222

Also good, the combination of recording interface (also your UCX) and ADI-2 Pro FS R BE (or ADI-2/4 Pro SE).

See some of my blog articles:
https://www.tonstudio-forum.de/blog/Ent … our-Setup/
https://www.tonstudio-forum.de/blog/Ent … ses-EN-DE/

The UFX III is interesting if you plan for a longer time (long-term investment) and might want to connect more devices, be it converters or preamps. It also offers more analog I/O where the number of ports doesn't shrink with higher sample rates (e.g., double speed). Another advantage, you can keep and further develop your setup in TotalMix FX. If you upgrade later you have to change everything and build-up everything newly.

Ideal is the combination of UFX III, ADI-2 Pro FS R BE (or ADI-2/4 Pro SE). It gives you enough headroom for a long time.
And it's nice that the UFX III got the converter of the older ADI-2 Pro FS reference converter.

Depends on your current and future demands, interests, budget .. as usual :-)

But all this doesn't solve your issue, that your PC might have some limitations in combination with Kontakt.
Where I am not sure whether we are talking about real limitations.
If you play with a master keyboard through a Kontakt instrument, then you do not have the full RTL.
The A/D part and transmission over USB is not applicable with such a use case.

+1!

Agreed! But only if you can use them of course. For example more headphones outs is an interesting feature, that I would never use. At least for now, again of course.

Thank you all very much for the responses. My fireface UCX (1) has not let me down yet, so I'll keep an eye on it. As is often the case, a replacement is not 'needed' but just nice to have. A week after I bought my new ucx, the new version came out. Maybe I should wait a while for the Fireface III, otherwise the same thing will happen to me again A UFX III seems great to me, but personally too much for me. If the latency doesn't matter with my old version, I'd better wait a while.

Ramses has a very thorough spreadsheet somewhere about the differences between all the interfaces, but I can't find it now.

I can't be of much help here, but around page 115 in the UFX3 manual it talks quite in depth about latency. In general, at 44.1 and using total mix to monitor, the round trip latency is .25ms. Other interfaces, like say an Apollo, around 2ms at 48 if I remember correctly.

Sound quality wise, I spent quite a bit of time listening to the UFX3 compared to the ADI pro2 fs, and I thought the conversion was close enough that I really didn't feel the need to use the adi pro for the DA side. How different are they? Good question, but I bet your room is doing a whole lot more to mess with the sound than the difference in those converters.

I'm currently using the adi pro 2 in a different room, a shed actually. So think pretty rural area, no room treatment other than a bunch of stuff creating some diffusion, but most importantly is behind the speakers (about 1 meter) I can open the "barn" doors (7' wide), so basically no reflections there, and from monitor speakers to the back was (which can also be opened with the barn doors is about 11 meters away. Ceiling comes to a point in the center at about 4 meters. Monitors are close, about 1 meter. Point being, I've never heard such detail and clarity as in this space. I can listen too a song and say "what that's really flat and compressed" or "that is super punchy" and I've never heard those differences from song to song. Or even YT videos, I can hear every little detail in the talking voices, and sometimes they sound good and sometimes they sound terrible, and I never would have heard that before. So is that the converter? Probably not, it's the room, and I don't think you can actually treat a room to do what I'm hearing in this space. Does the ADI help? probably, but I don't think that's the biggest factor.

My other main recording/playing space with the UFX3 is fairly well treated and I've lucked out and have it pretty flat, like within 6db down at couple points in the low end. That area is about 20x40 feet with a low basement ceiling. I have another area which is almost the exact dimensions of Eric Valentines control room, except the ceiling is lower. It's the same golden triangle as Eric's, and it was completely problematic as it was for him (see his 29 videos dealing with it). I have about $2k worth of the nicer auralex stuff in there and was still a huge problem. I ended up putting the monitors about 1/3 into the room facing to the short wall and it sounded good and I could make it work, but it was a disaster with the monitors anywhere else in the room facing any other way.

The point I'm trying to make is the room has such an impact on your monitoring that the difference between these converters is pretty much a moot point from the sonics side. They are all very good.