I actually did some testing with this, and a lot of factors come into play. I have an Antech 900, for the record, along with all low-speed Yate Loons and the stock 200mm top fan. That means 2 front intake, 1 side fan (this is temporarily a OMG low speed Coolmaster since I broke the Yate that was there), 1 rear exhaust fan, plus the top exhaust fan.
I found -- with video cards that vent mostly inside the case -- most temps were within 1*C difference on both video card and CPU so long as I kept the two front fans on intake. All other fans exhausting, all other fans intake, all other fans mixed...didn't matter. *BUT*, when I changed to a different video card -- also venting inside the case -- by flipping the side fan to exhaust I actually decreased the card temperature by 2-3*C. The video card was venting inside the case but the side fan was directly located next to the vents on the video card. By directly sucking out a lot of this hot air, it helped.
Moral of the story: get a sheet of paper, a good thermal benchmark, and try different configs. Different cases can lead to different needs as well.
After seeing that artical JBDan mentioned, I might have to try my test again with the front fans disabled. It doesn't seem to make sense with no fresh air being actively vented into the case, but hey, who knows
@OptimumSlinky: See above. I noticed a drop when I switched my side fan to exhaust, but the location might be pretty key. If it's too far down it might acutally interrupt the video card cooler's ability to pull in cooler air. In that case you migth want it to vent inward.