I just went through this question myself.
If you go used, go for the cheapest model. The differences are small. I missed out on some good K-1 deals but ended up getting a K-1 Mark II.
If you look at the Pentax sample gallery at dpreview, you can see the differences in processing. Sharpness initially looks better on the original but if you look at the CORNERS (the red watercolor brush), you will see that the K-1 Mark II is sharper. This means there is a subtle difference in focus points and with the crazy pixel peeping, you can see it.
Next, if you look at PentaxForums’s review and look at the star eater page, you can download the full res photos. Here you can see that there are fewer stars in the K-1 Mark II but that this is probably a focusing error to since the stars are not as pinpoint. The background sky is also better on the K-1 Mark II.
Pentax has stated in some cases that the accelerator may be worse, but I think stacked astrophotography may be the only time that’s happening. I have not seen enough K1 Mark II astrophotography pictures to really tell what is happening (you also need advanced optics on a tracking mount to be confident that you are not just seeing thermal shift between lenses).
PENTAX K-1 Mark II????????K-1??????????DPReview?????????? - ?????
Is a good comparison.
——-
Last, as you look at the noise, remember that different companies use different measurements. Pentax is pretty honest. At ISO 100 both the K1-Mark II and Panasonic S1R have a 1/40s
Go to ISO 25600 and you see under JPEG that Pentax is 1/6400 but 1/5000 for the S1 and A73. There are some weird differences that way and the RAWs are 1/5000. Makes you wonder why they don’t shoot RAW+JPEG?
At DxOMark you can see how the A73, S1 compared to the K1 for ISO and low light. But remember that in good light, you get 36MP