Veritas quant questions are sometimes slightly more complex than those on the real test, and the verbal can be trickier somehow. At the same time, most people score pretty much the same on the real test as they do on Veritas tests. That being said, occasionally people score significantly lower on Veritas tests than they do on the real GMAT.
Many Manhattan quant questions are so constructed that getting to the right answer to them takes more steps than one needs to get right answers to questions on the actual test. Real GMAT quant questions tend to be designed to test vision. In other words, often just seeing a key thing is most of what you need to do in order to get them right. Many Manhattan quant questions often require doing more actual math. Also, some Manhattan quant questions are simply more challenging than quant questions on the actual GMAT tend to be.
Manhattan verbal is often pretty close to official verbal, but because of weaknesses in some of the questions, people taking a Manhattan verbal section may get more wrong answers than they get when taking the real GMAT.
So, because of the combination of more math heavy quant questions, challenging quant questions and some flawed verbal questions, people tend to score a little lower on Manhattan tests than they do on the real GMAT.
GMAT Prep? I can say that from what I have seen and heard, SC on the real test probably has become more logic based than it is in GMAT Prep. So in a sense that change makes SC on the real test easier in that maybe there are not on the real test as many somewhat arcane idioms to deal with as there are in GMAT Prep. Another way to put it is this way. I like nice logical, hackable SC questions and the last time I took the test, I felt that I liked the SC questions I saw more than I like those in GMAT Prep. Is quant a little tougher in GMAT Prep than in the real test? Maaaaaybe slightly, but maybe not. I guess probably GMAT Prep quant is pretty much the same as quant on the real test.