Taking the ratio of men's world records, we have approximately:
5 seconds in a 10k <-> 23 seconds in a marathon
Or if you prefer,
1 minute in a 10k <-> 4:36 in a marathon
In other words, all else equal we would two runners who are 5 seconds apart in a 10k to be 23 seconds apart in a marathon. Of course, all else is not equal, but this isn't a terrible way to compare runners who prioritize the same event (i.e. Kiptum with other marathoners).
By this, we would expect Kiptum's 10k time to be:
10 seconds faster than a 2:01:21 guy
20 seconds faster than a 2:02:07 guy
30 seconds faster than a 2:02:53 guy
40 seconds faster than a 2:03:39 guy
50 seconds faster than a 2:04:25 guy
1 minute faster than a 2:05:11 guy
2 minutes faster than a 2:09:47 guy
3 minutes faster than a 2:14:23 guy
4 minutes faster than a 2:18:59 guy
Again, this isn't perfect, and Kiptum is unique by any standard. But there's a tendency to estimate HOW MUCH BETTER Kiptum is than other (non-Kipchoge) marathoners.
My personal intuition is that most 2:05-low guys can almost certainly run 27-low with a few months of specific training [certainly, the marathon time is stronger; I am talking about a marathoner moving down, the way Kiptum would]. Similarly, if you can break 2:10, you can probably break 28, or otherwise come really close. 2:14 guys can run 29-low, and 2:18 guys can run 30-low. On the other end, anyone breaking 2:03 would be well into the 26's. And I don't see any of these are being substantially easier/harder than the others.
All of this corresponds to a 26-low 10k time for a "typical" 2:00:35 guy given a few months specific training. But maybe your intuition is different than mine. Or maybe you believe that Kiptum is so unique that the normal rules don't apply (I am always skeptical of these arguments). Just remember that we're dealing with a guy who wins majors by MULTIPLES MINUTES every time he runs--it wouldn't be surprising if he can drop a massive 10k time as well.