I would go with an electromagnet simply because the physics of how electromagnets work are simpler and can be easily calculated (sounds counterintuitive, but true).

As for the bismuth, I thought you already had a sample - if you don't, then getting it is going to be a problem. I would say to get as little of it as possible but which you can still measure, maybe a gram or something like that, or less if you have a scale that can measure smaller amounts than that. The first reason I say this is because bismuth is a rare element, only about twice as common as gold, which means that it will be very expensive. Secondly, even if you are using a very weak electromagnet, you won't have as big a force to deal with, and there is less chance that you will break your magnet.

Unfortunately, I do not know where you can purchase electromagnets suitable for such an experiment - nor would I really be qualified to tell you what would be a good one to get (again, I'm a biologist). If catalogs for toys and such (e.g. Edmund Scientific) sell electromagnets, and they might, then I would imagine that they would be powerful that you could use them to push a gram of bismuth around - heck, if you're a good enough mechanic, then you could even disassemble an electric toy truck and use its motor, since an electric motor is an electromagnet - and that motor also is plenty strong enough to handle a gram of bismuth. But honestly, I don't know what is out there. You may need to get yourself a ferromagnet - although a decent ferromagnet should still be adequate to the task.

I suppose the first question is what your hypothesis is - do you simply want to demonstrate that you can exert a repulsive force on a piece of bismuth using a magnet, or do you actually want to levitate the thing? If you want to levitate the thing, then you're going to need to do some engineering, and a much more powerful magnet.