Pentax has it's strengths and weaknesses in AF, just like many other brands.
It seems common to assume AF only has to be is 'fast' to be good. But often other dimensions of AF performance are missing from this common measure of 'good'.
A few things that spring to mind are the accuracy of camera AF performance (with related issues of how
consistently accurate the AF is), and the working range of the AF, in particular it's light sensitivity. If the AF can't work in low-light, a whole range of shooting scenarios are off-limits, and the calculus of what 'good' AF performance means needs to reflect that.
On the 'accuracy' variable, Pentax has usually tracked well here from various testing sites (if memory serves, demonstrated by tests from France 'Chasseur d'Images' & 'DigitalFotoNetz' in Germany) where the measure is reliably producing sharp images. But there is a big caveat, of course, to any claims of superior
shot-to-shot AF accuracy in any PDAF system, where accuracy depends on, amongst other things, normal camera/lens variation.
On the matter of low-light AF, it's notable that Pentax has consistently led other DSLR brands with the low-light sensitivity of it's camera AF: the K-5II came with -3EV AF sensitivity in 2012, whilst the first Nikon to meet that specification (the FF D750) came out in 2014, and the first Nikon APS-C to match -3EV (the D7200) only came out in 2015. Similarly with Canon - the FF 6D did -3 EV in 2012, but no other Canon until the APS-C 7D2 in 2014 was provided with the same sensitivity. And despite the 645Z/K-3/K-3II/K-S2/K-70 etc coming along, no Sony got anywhere near -3EV EV AF sensitivity until the A7S was released in 2014.
TL; DR - AF is a huge can of worms. Pentax does OK, depending how you look at it.