I think there might be a bug with the test or at least the engine. I don't think it always took my settings when I tried to enable aa and af. Anyway I adjusted my fps to what now I think is a true representation. 1600x1200 seems right to me and I'm still showing a rather large discrepancy at 1280x960. Something is definately going on with these benchmarks and I'm clearly not the only one that thinks so.