Actually, it means that out of 100 representative students his age in the US, 99 would score below him. He would be equivalent to the 100th student.
Otherwise, I would agree with others above.
It is unlikely that you will obtain more information from the CogAT than you already have from the comprehensive evaluation (IQ, achievement, etc.) conducted for his IEP.
Changes in raw scores do not always have an obvious correspondence to changes in percentile, as it depends on the standardization process, where the change occurs in the curve, etc. Given your existing, much richer, data set from the IEP eval, I would not expend too much time and energy thinking about the CogAT.
