at this point, i would say wind every other tooth for a concentrated pole winding. its half the complexity, and you can probably get more copper in it than winding every tooth. also there is less chance of a short between two coils or between two phases.
because you have to start over with cutting slots, the number doesn't really matter. pick the number of magnets you can place on the rotor for the most number of magnets, ideally the ones you already have. if they are 10mm wide, then go with that number, 15, etc, 20mm, etc. or if you have imperial magnets, 3/8th, 1/2", 3/4" or 1" wide magnets. pack them close together and use that number.
once you have the maximum number of magnets on the rotor for a practical and available width of magnet (that is, use what you already have, and or use the magnets that are cheaper per cubic inch), then choose the slot count that makes sense.
if you go back to that site i link to repeatedly, you need to choose a pole count and a slot count that makes sense. winding factor should be as high as possible. so should the least common multiple, higher = less cogging torque. also, avoid the choices that force you to use a 2 layer winding. on that website, 2 layers for concentrated pole motors means you have to wind every tooth.
I still tend to think that the magnet length to width should be approximately 1:1. meaning if your lamination stack is 20mm deep, then use 20mm wide magnets. this dimentionless ratio seems to hold across a lot of different types of machines, but we see exceptions everywhere.
for example, the 100MW direct drive hydro turbine generators that are 20meter/60foot diameter. the poles are a foot wide, all 120 of them, (or whatever number it is) and 4 feet deep/thick. perhaps in an ideal world the generator might be made to be 40meter/120foot diameter, and the poles 2 feet wide and 2 feet deep. the air gap would be the same 4 square feet, but the electrical length of the copper windings would be reduced from 10 feet to 8 feet. diminishing returns at this point. clearly they can't make the generator that large diameter, so you make it deeper instead. the weight would be the same, but the air gap velocity would be double, so you would multiply the windage losses by 8, but divide by two for the reduction in depth. so windage losses would be 4 times higher, for 20% less copper losses.
so, i'd use whatever magnets you already have. i don't think the pole count is significant until you discover what your eddy and hysteresis losses actually are.