This used to be true on the classic GMAT, where your performance on each section (except for IR, which would simply provide the overall % of counted questions answered correctly) was broken into quartiles. Even though at one time the GMAT was approximately 25% uncounted experimental questions, GMAC would simply omit any experimental questions from the ESR, as if they never existed--and this worked for the most part, because the old ESR didn't indicate any actual question numbers.
Lucikly, we were still able to deduce the total number of experimental questions on the classic GMAT via the denominator method. For example, the old, classic GMAT after the addition of the Integrated Reasoning section (June 2012 - April 2018) featured a whopping
23 experimentals out of 90 questions total (25.6%), and the shorter, most recent classic GMAT version (April 2018 - January 2024) featured
12 experimentals out of 79 questions total (15.2%).
However, the new Focus ESRs show the results of all 20 DI questions, all 21 Quant questions, and all 23 Verbal questions, so there is no doubt in my mind that
some of these questions are in fact experimental and don't count toward your score--which, as Ben has said, is supported by a student scoring 100% on RC despite answering 1 RC question incorrectly.
In fact, I believe it's only a matter of time before a test-taker scores a perfect 90/90 with 1 or more questions answered incorrectly, and that the only reason why we haven't yet seen this is 1) there is a small sample size of Focus ESRs publicly available so far, and 2) a high-scoring GMAT student who answers all the hard, counted questions in a section correctly is unlikely to answer a medium-difficulty (on average) experimental question incorrectly.
If we assume roughly 15% experimental questions on the GMAT Focus, this would equal
9 or 10 experimentals out of the 64 questions total.
https://gmatclub.com/forum/how-many-experimental-questions-on-the-gmat-and-is-it-predictable-222648.html