In Marvin 5.11.1, I get the following:
On the website, I get:
Notice the difference...
Additionally, on the website, the N and O 'groups' (indicated) do not have a value...
Hence, also the species distribution is different.
Which one is the correct one? I assume the Marvin 5.11.1 one?
The Website pKa calcualtor try to follow the "golden mean" even in case of the minor tauotmeric form had submitted for calculating pKa values. It means that only the most interesting pKa values are reported . Usually this strategy provides reasonable pKa result, providing that the submitted molecule is not too far from the dominant tauotmeric form.
On the other side , in MarvinView the pKa calcualtion is "liberal" enough. User has a chance to choose options and decide what to do with the structure before of submittion to the pKa calculator.
For example incase of iopromide the best choice of options seems to take the "consider tauotmerization" option.