The picture here is the insides of a cheap antenna I bought (should be like the ebay link given in post 59). The pic also shows the extra bit I soldered onto the tip.
The antenna was advertised as 868MHz, but when I checked with my N1201SA antenna analyser, it resonated more at 915MHz. I added the extra length, and it brought down the resonant frequency to 868Mhz (happy days!). Whilst I can’t give exact values for performance, I can say the RSSI did improve, so for sure, the antenna is still radiating OK etc.
Be aware that for resonance, lengths that are measured can be shorter than that predicted by the simple formula “C =freq x wavelength” This is because this simple formula does not include some other calculation factors that are required to be included in the formula. These extra parts can be too compicated and variable to include for this sort of antenna - Trial and error in cutting the length seems to adjust fine!).
In the end, the presence of the sheath reduces the overall reactive part of the impedance. In other words, you must put the plastic sheath back on. If you don’t the impedance is out.
Have a look at the following link (post 432) for more info about what I did
In the end, I decided that I wasn’t going to get a significant performance boost using a different antenna. However, I did find that if this antenna had been placed high up (with direct line of sight) in free space (and not inside a building on the ground floor), the performace boost was about 1000 times better (i.e. from -71 db to -41 db) .
When you are looking at antenna options, you won’t get one (outside of yagi) that is practically (and within reasonable cost/size considerations) more than 4 times better than a standard correctly resonant dipole (such as this cheap antenna when tweaked) . If you do, the polar pattern becomes compromised e.g. like a directional yagi.
This is the point - 1000 X performance increase is a lot more than 4 - and you will not get an antenna that gives the sort of boost that placement does !!
Anyway, it sounds like your gateway antenna is a bit iffy or its placement is wrong (near metal etc?)
When I checked the antenna of my TTN gateway, I found it was exactly the same as these cheap antenna. I made the change as above it definately improved the situation.
If its good enough for TTN gateways, its good enough for me (possibly the wrong assumption 
)
Also, be aware that even though you can stick a gateway against window glass, some modern glass has metallic coatings which severely attenutate the signals.
Theres no getting around it - place high up and free for max performance. There are other reasons why it should be even higher, mainly to help the resulting polar pattern NOT being skewed up away from a downwards direction (where your devices probably are!).
Also bear in mind that you do not need a ground plane with these cheap antennae
I know it doesn’t answer the question of “which antenna is better”, but I hope this info assists you with another direction for improvements - In that you should check “placement” first if you want long distance. If the antenna is for the gateway, then outside is better, so you should look for an outside antenna which is placed somewhere appropriate.
If it is for a node, then consider cheap antennae (if you can get ones with the correct lengths!), or a straight up wire at the correct length.
N.B.It does look that the ones you have pictured are not resonant for 868MHz
Hope it helps 