They're looking at 2 slightly different things. One of the problems with gifted programs, especially in public schools, is that they're not really calibrated to differences between gifted kids. And more importantly, they're not really advanced learning environments.
I remember how morose my son was when, in 4th grade, a well-meaning substitute teacher gave him a booklet of sixth grade maths to complete. His regular teachers would just let him ponder the maths videos that he watched at home and at that time, he was interested in how infinite sets of numbers relate to each other. Out of politeness, he completed her booklet and then he asked her not to give him any more exercises unless she actually evaluated the level of his maths abilities and thereafter he was left (happily) to his own devices again. Objectively, that year, the school entered him in a sixth grade level maths competition and, without any effort into preparation, his score was within the top 1% of sixth graders nationally and in the following two years of the competition, he achieved perfect scores.
In this age of digital technology and highly connected information networks, the ‘best’ form of gifted education in primary schools, ill-equiped to teach the very highly gifted, may simply be to give them access to resources for self-directed learning.
Edited to add: I will forever be indebted to the Head of the Maths Dept of our local high school who realised at the start of 7th grade that he had already mastered most of the Yr 7-10 curriculum but that there were a few gaps. She consulted teachers at other schools and decided that the best approach was to use an online maths program which included pre-topic assessments, so that students could skip topics they had already mastered, but ensured that they covered the entire curriculum. If my memory serves correctly, he skipped all of the Yr 7 topics, more than 90% of Yr 8, 70-80% of Yr 9 and more than half of Yr 10 and he was able to cover all of the outstanding topics in the first six months of high school.