AD633 , higher orders ?

Hi all, I wanted to use either dy/dx or the double angle formula to add overtones to an audio signal ( specifically with advancing the slew rate in a manner proportional to frequency, not limited by an RC constant).  The AD633 looks great, and even will produce cos(x) from sin(x) with few additional components.  What I wanted to ask is whether the AD633 can take the n of x^n beyond 3 or 4. Specifically, if y=sin(x) + k*cos(x)^n, what is a reasonable upper limit to n, assuming the gains of the successive iterations are controlled?

Any help appreciated.  For a visual, 

Thanks, 

George