The gauge is one of many choices you have, maybe not the most relevant. At very high frequencies, the skin effect is so prominent that significant variations in gauge make not a great difference, except in the rigidity of the resulting coil, ease of winding, and possible winding pitch. Take a look at this
magnific page: for a 15mm diameter coil, with 5 turns and 8mm length, using 18AWG gives 339nH with Q=687 at 100MHz, while the same coil with 22AWG gives 368nH with Q=699. The series resistance of the 22AWG coil is just 9% higher, while the wire diameter is 47% less. The resulting Q and inductance are in fact higher. I spent a lot of time winding my own air coils for a similar project some years ago, and I wound up favoring the
greatest lowest (greater diameter) gauge possible for rigidity and control, using big screws as formers, and approaching the desired inductance with the page linked above. It worked out rather well, considering my huge inexperience.