"Academically the child should demonstrate skill levels consistent
with the average performance level of the students in the grade
desired."
There is no mention of >95%
This echoes sanne's question
upthread, as to the average performance on the end-of-4th-grade math assessment.
Unfortunately, as mentioned in this
old post, it is not unheard of for schools to create more stringent criteria for students
accelerating into a grade, than for students
passing into a grade. This could be challenged when advocating for your child. Did you see
this post? Possibly with a bit of research, you may locate the source information in Susan Assouline's work and be able to utilize it to advocate for a cut score at 80% rather than 95%.
While I tend to agree with ruazkaz
upthread, in saying that a child who has done well studying AoPS ought to do quite well on an end-of-4th-grade math assessment, I would also add that it depends on the quality of the assessment itself. For example, does the assessment check for concepts, strategies, and execution... or might it test for some esoteric knowledge that only a child who has studied a particular curriculum would tend to know?