Not much there to go on, but that said the Wikipedia article is substantially outdated. A more recent description of field experiments can be found at:
http://www.jhuapl.edu/techdige....
There are plenty of papers paywalled at the SPIE site as well if you'd like to get a better feel on what the state of the art is post the 2012 experiment described in the linked article.
All that said, the environment is your enemy as you go up in frequency - things like clouds, rain (but not always), fog (again, not always), and the pigeon that someone mentioned previously can break the link, but proper design of the modems can get through some of that. That said "some of that" will not get you 5 9's link reliability over all weather condition. It's not just optical; the mm-wave stuff being thrown around for 5G systems has many of the same problems - heck, certain bands of high frequency (say around 60 GHz) get soaked up by oxygen at incredible rates.
Regarding data rates, RF systems will struggle to beat FSOC. Well designed FSOC systems leverage technologies used in fiber communications; I've worked on air-to-ground links that exceeded 80 Gbps over a decade ago, it would be trivial to double (or more!) that rate.
Another consideration is cost - FSOC systems in general utilize pretty sophisticated optical systems that are effectively your antenna. The modem cost can be driven to par pretty easily, but RF antennas are generally much lower cost than optics.
But you don't have to fight for spectrum, which is a major advantage. The FCC and I imagine nearly every other country's spectrum regulatory agency do not regulate the near IR wavelengths FSOC systems run at. There are safety issues that need to be taken into account (eye damage), but those are well defined and don't generally represent that much of a problem.