coplanar waveguide vs microstrip for gnss and cellular? I read on internet microstrip is better for signals <30Ghz but some gps manufacturer (Delorme) said that coplanar is most efficient so it's kinda confusing me. any advices?
'Better' and 'most efficient' are not words that are useful to use.
Both microstrip and coplanar have their own advantages and disadvantages. However, often it's the integration with the significant ICs on the board that's the key factor. Some ICs have pinouts that invite connection directly to coplanar (with a GND-sig-GND set of pins) or microstrip (with a ground pad and a single signal pin).
Coplanar makes it easier to use shunt components, they just go straight across the gap without the need for vias. However, the need for balance between the two grounds needs them tacked to each other sufficiently often. The top surface grounds can crowd the available real estate.
Coplanar-with-ground is a hybrid of the two which can be used to transition between them.
There's little significant difference between the attenuation constant or dispersion of the two.
Neil's answer is good, but I'll add a couple more points that might affect your choice of geometry
There are good empirical formulas available (and online calculators that implement those formulas) for choosing microstrip trace width to achieve a desired characteristic impedance. The online calculators for coplanar are more likely to disagree with each other, leaving you wondering what to do.
If you have a very thick substrate (for example, a 2-layer board on 1.6-mm thick FR-4), microstrip trace geometry might be inconveniently wide; coplanar can be designed to give a narrower trace (but of course the ground areas also require some board area)