Holy ###.
My first response, before even looking at the screenshots, was: It isn't enough to look at the html code. That's simply what is sent to your browser. You need to know how the html was created in the first place, whether hand-rolled or made by custom php or a CMS.
My second response, after a cursory glance at the screenshots, was: Oh, I get it, all the non-ASCII characters are getting converted to decimal entities. (Interestingly there was a post just a few days ago leading to the discovery that some mobiles don't "do" decimal entities; it has to be hexadecimal.)
In fact it's worse: about half of the ordinary letters are being rendered as decimal code for absolutely no reason. I really, really hope this isn't happening in the server, as it means that about every other letter takes up seven times as many bytes as it needs to.
The encoding is random, rather than consistent: here an "o", there a "о" and similar.
One of your two screenshots shows the text as-is-- except for the somewhat glaring issue of <br> where <p></p> seems to be warranted. So what's wrong with starting with that version?
In any case, there's no earthly reason why ordinary ASCII characters would be converted into decimal entities. You may need to lean a little harder on your writer.
