The short answer is: U Mβ is not higher than U Mα. It is spectroscopic nuance of P10 (90 % of Ar) based gas proportional counters. (If You will look to EDS You will see that U Ma on Uranium standard is higher than U Mb). It happens that Ar K absorption edge is in between U Ma and U Mb. There are few overlapping processes which makes that difference of measurement present.
First U Ma is at lower energy than Ar K edge, and is better absorbed by Ar K shell, thus the X-ray of that energy ionizes less of Ar (this is actually questionable for me too, where the absorbed energy go? does P10 emit non-ionizating radiation?).
Second process is that higher energy than Ar K shell will cause secondary xray fluorescence of Ar, which will creat additional ionisation of gas chamber, which is known as "Ar escape peak" in PHA. If You would filter Argon escape peak at PHA, and would do WDS scan You would see, that U Ma is actually higher than U Mb.
Last but not least, You should not look only to absolute height of peaks, but for intensity comparison you should normalize peak height with background height (I_peak/I_bkgd), without any PHA filtering You will see that actually even on P10 gas proportional counters U Ma is more intense peak than U Mb, as Bremstrahlung background is buffed-up at lower theta (higher energy) than Ar K edge too.