A little? The difference between those scores is literally less then half of a percentage point. Which probably converts to what... an eighth of a frame on average improvement? Unfortunately, even the perfect GPU graphical benchmark will always have some variance between runs. I've seen a whopping 2% difference just by running a benchmark twice in a row without changing any conditions. Still not much, but more than 4x larger difference then your runs. If you desire results that can come as close as possible to showing what improvement exist, if any, you will need to take multiple runs and calculate an average. 3-5 runs under the same conditions is typical. If you want to find out if there really IS a 0.04% improvement I'd say you are simply SOL. That kind of precision just aint happening.