Saying that you gave the first proof of Fermat's Last Theorem is too good to be true. (I.e., it is literally false, since the result was resolved 20 years ago by someone else.) Saying that you are more talented at physics than Einstein and Feynman combined is technically possible but strains credulity to the point that it would have a strong negative effect if not backed up by some truly remarkable facts. However, saying that you are in the top 1% compared to all students in the last five years is obviously not too good to be true: it must be true of at least one student. (These questions are often muddied by not being precise enough about the cohort being compared, and you should know that admissions committees interpret them with a grain of salt.)
When I was involved with PhD admissions in the UGA math department, each year I saw several applications in which the recommender gave the applicant top marks in every category. When this happened I didn't say "Ridiculous!" but instead looked carefully at the rest of the application. It may be that I conclude that the recommender is a bit naive and/or hasn't seen as good students as I have...but that still might mean that the student's application is quite strong. In general top marks are good things, not red flags.
To my mind the fact that two of your recommenders showed you the letters is much more of a red flag than the top ratings. The strongest letters of recommendation often contain confidential information that would not be suitable for the candidate to read (e.g. comparisons to other named people). If such information doesn't appear then there is nothing inherently wrong with the practice, but nevertheless it does not inspire my confidence.
I guess if you are looking at the recommendation letters you have a chance to evaluate their suitability (which you can use in a future year; it is awkward and perhaps even ethically suspect to withdraw a recommendation letter after reading it). At least in US graduate applications, good letters are about a lot more than the slightly silly ratings. They also contain several paragraphs of text, usually occupying at least the better part of a page. If someone gives you absolutely top marks and then writes little or nothing to back them up, they look quite lazy. Though that does not specifically reflect on you, it certainly doesn't help your application either.