Interesting. Got a resource handy regarding the costs of enriching past 4%? As you stated earlier, the difficulty (and presumably cost) drops off at the 20% mark, so I'm curious that if you take the enrichment all the way up to 90-95% the much longer "burn" time in the reactor (ie. the cost of all the refueling shut downs that would be avoided over the span of 20-30 years) would counterbalance the initial additional cost of going from 4 to 20%.
Also, how much plutonium would be produced and could the spent fuel be directly reprocessed into MOX for subsequent re-use?
ETA: Also is it possible to create a dual-fuel breeder/MOX reactor that burns U-238, and then can have the spent rods processed and plugged back in, in order to get the most "bang-for-the-buck"?
ETA Again: I just want to state that I'm under no illusions that Iran is going for the bomb. Just wanted to play devil's advocate a bit, plus I'm genuinely curious about nuke power. I was slated to go to nuke school but got my orders canceled at MEPS when we found out I've got a slight red/green color vision deficit. 
Here you go for the formulae:
http://en.wikipedia.org/wiki/Enriched_uranium#Separative_work_unitAnd the results:
To turn natural into 4% requires 6.6 SEU to yield 1kg of 4% for every 7.6kg put in.
Natural to 90% is ~230 SEU to make 1kg of 90%
Natural to 20% is ~46 SEU to make 1kg of 20%
20% to 90% is ~20 SEU to make 1kg of 20%. (since this requires 4.5kg of 20% as feed, the total is ~210SEU)
What this means is it costs 40x as much to make 90% as it does normal 4% fuel.
The other MORE IMPORTANT fact is to make 1kg of 90%, making the 4.5kg of 20% means that YOU HAVE ALREADY EXPENDED 90% of the effort required to go all the way.
So put it this way, let's say you need 100kg of 90% for your evil project. You would need about 18,000kg of natural uranium.
So to avoid "being the bad guy" until it's too late, you build centrifuges capable of 20,000 SEU/year, and at first make 450kg of 20% "research reactor fuel"... That takes a whole year, but you are still "a nice country, not making a bomb".
Now, secretly, you simply TURN SOME VALVES ELECTRONICALLY on your centrifuge plant, which unless you know which way the valves are pointing, and can crawl all over the schematic and physical nature of the pant to see what pipes go where no one can tell you did, and VOILA, that same facility can turn your 450kg of "LEU" 20% into 100kg of 90% bomb grade material in LESS THAN 5 WEEKS!!!
THAT is why, to quote our current VP, it is a "big f-ing deal"--any enrichment facility capable of creating hundreds to thousands of kilograms of 20%/year is EASILY CAPABLE of turning that into bomb material in a very short time, with little way of outsiders detecting the change...or even insiders, unless they assay the output.
If you were just making PWR 4% fuel, the above facility would be able to make 2400kg of PWR fuel per year..,enough to fuel a VERY large (for lets say, Iran) power plant, or make enough "medical isotopes" for many many bombs.
Like I said, the only reason to make more than a few 100's of Kg of 20% is to make a bomb.
As for breeders, for power, LEU startup/MOX burn breeders are THE most efficient way of using uranium to make power (or plutonium for that matter).
Sorry to hear about not getting nuke school, but a red/green deficit even if minor is a disqualifier since there are many times green is "okay" and red is "oh god, oh god, we're all gonna die"
