In Appendix A of Harold Gans' primer, Rabbi Dovid Lichtman examines the issue of the accuracy of our current Torah. He shows, via historical records, meticulous practices of Torah scribes, and both modern and ancient comparisons of Torah scrolls separated by time and geography, that our current Torah has at most 9 to 12 letters that can not be resolved by the usual methods of comparison used by the Masorites. This is out of a total of 304,805 letters in the Torah.

I prefer not to take sides on the arguments surrounding the text's accuracy, and there is no need to from a codes perspective, as follows:

Foreknowledge of our current text

If we truly allow ourselves to believe that the Encoder is infinitely omnipotent, then we understand that this includes His ability to "see" our current text and encode it in a single timeless "instant," when the original Torah was first conceived. This may sound impossible until we can look at a concrete example. The Coder's apparent knowledge of minute details of the eventual translation of the Torah to Greek (the Septuagint), is a case in point (see the code examples here).

Certificate of Authenticity

It is significant that the vast majority of Torahs in synagogues throughout the world today agree letter for letter with each other, and with the edition by Koren Publishing which is used in all disciplined codes research. If anything, the existence of codes provides a kind of "Certificate of Authenticity" (an approval, if you will) for our current edition, without any implications about how this text was transmitted over the centuries.

Backwards Logic

Actually, the whole accuracy argument is backwards: rather than alleged inaccuracy implying no codes, the overwhelming statistical evidence for codes imply amazing accuracy.

Unaffected, localized codes

In addition to these points, there is also the fact many Torah Codes, among them the most significant ones, span fractions of a per cent of the text. Many of these "localized" codes would be wholly unaffected by minor text variations (many such codes would remain intact even by a worst
case scenario of extensive variations).

Update summarizing the relevant Talmudic passages and what we know of the history of scribal transmissions

Please refer to this article, under "Harold Gans reponds".