There is no easy way for you to know the flux density in the core except by the voltage you get at a certain speed.
The first place to saturate is the core teeth, when this happens you still get some increase in volts but it needs a lot more magnet to produce a few more volts.
I haven't done anything with motor conversions but from experience with wound field machines, I suspect some of the conversions must be near saturation at least in the teeth.
I am inclined to suspect that with curved neos that fit the rotor and stator curvature and covering 2/3 of the pole pitch, you will be close to saturation with magnets 1/4" thick.
Using rectangular blocks which don't fit the curvature and leave large air gaps may require a lot more thickness.
From the output point of view, taking the teeth into saturation is good, but it will give more iron loss. Like all things there is probably an optimum between low speed performance and maximum output.
Flux