I can tell you why they are different, because a rich-text token can be used in either an email or landing page.
In a landing page, there is only one link option and that is to make it untracked. The way you do that is actually different than in an email and it is done by adding class="mchNoDecorate" to your <a> tags.
In an email, there are two options for links. The first is to make it untracked and the way that is done is by adding class="mktNoTrack" to your <a> tags. The second option is to either include or not include the mkt_tok query string parameter in the destination page. mkt_tok is only applicable to email links since that is what helps Munchkin associate the lead with the cookie.
So, if we wanted to have the behavior in rich-text program tokens, we would have to have "Don't track" add both classes behind the scenes and change the wording of the second option to be "Don't include mkt_tok for Emails".
If you think that is better, file an idea and we can consider adding it. This is just why the experience is different today in the different places.