We try to mimic the real test as much as we can! The standard error on our test is +/- 50 points (compared to +/- 30 points on the official test). This means that about 2/3 of the time, our test will predict the GMAT score within 50 points. (And the official test predicts your score within 30 points - it sounds funny to say the official score predicts your official test score, but the idea is that you wouldn't necessarily get exactly the same score if you took it two days in a row. There's a standard deviation to the test.)

My students often tell me they think our quant section is harder than the real test's quant. One of the primary reasons is that our test doesn't include experimental questions, so you don't get a few lower-level questions that give you a "brain break" in the middle of the test.

On verbal, I hear pretty equally from people who think it's easier and harder, which usually just boils down to individual perception - it can be easier or harder depending upon your individual strengths and weaknesses and the questions you happen to get.