I am just surprised that they vary so much depending what model is been
used.

Really? Mostly, I think they are within a degree.


What model is the best? And also it seems that by the poles the calculation
are somewhat wrong, is it just my imagination?

The compass is pretty much unusable near the magnetic poles, mostly
because of dip and also because the variation varies so rapidly.

Either the Defense department's WMM or International Geomagnetic
Reference Field models should be good enough for navigational
purposes. Mostly, you'd choose one to be consistent with some other
data.

Jepp used to use WMM, but when I asked them recently, they said they
were using IGRF. Most of the world's airlines use Jeppesen data.

Ed