Answer after edit:
If you are getting perfect scores on standardized tests, I wouldn't doubt the general horsepower. Perhaps there is some very specific aspect of mental processing that enables you to perform well on GMA tests, but not on brain teasers. And perhaps even there are fields that benefit from fast brain tease ability. But this is a subtlety, among many when you look at what to do with your life (which has no Euclidean answer).
However, the comment on picking right problem is very interesting. One of the most important aspects (very underrated) is having good intuitions about which problems are high likelihood of success. Lot of sad stories from grad students who spin their wheel on "a problem the boss gave me". Conversely I picked something I knew would work, sounded cool, and was cool; and project selection made my time so much more calm. I think almost every field has cool problems and that when you get into it you will find them, especially if you think identifying good problems is a skill of yours. Specifics will come from getting into the field though.
Personally I would go towards whatever interests you. One small advice I would give is not to underestimate the experimental or applied sciences. I was the classic "math science" type who wasn't perfect at a lab notebook and test tube cleanliness (and was scared by subjects involving difficult experiments or apparatus building...figured I was a pencil/paper type). However, the truth is that a lot of what determines success in those fields at the real research level (as opposed to school courses) is picking projects and ability to read and interpret the literature. And they are a lot of fun...
Answer prior to question edit:
Your question is rational and candid. We all have to look for comparative advantage.
There is a huge literature showing a correlation between IQ and performance in mentally complex jobs. See, for example, meta studies by Hunter and Schmidt.* (It is the strongest statistical relationship in nonpathological psychology.)
Of course, nothing is 100% as a predictor, but in general there is a certain amount of smarts needed to handle some areas like grad physics (Jackson E&M) and grad math (God knows). And having more seems to be better even for those who have the minimum in those topics.
Again, it is not a 100% predictor. Neither is speed or strength or size for certain sports. But it is a strong factor. Consider that there have been only 25 NBA players in the history of the league who were 5-9 or less. Conversely if you meet a 7 foot man, there is a 1/6th chance he has or will play in the league. So obviously specific traits (talents) can influence performance at high levels.
Note that there is some sorting of and probably different floors by academic topic. With the more mathematical topics needing more horsepower and the more applied ones less. See XKCD: https://xkcd.com/435/ However, be of good cheer. The good part is the more applied ones are more interesting--more hair on the ball, more direct connection to industry/society. Or so I tell myself to feel better for not solving Fermat's Last Thereom. ;-)
*Link: http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.172.1733&rep=rep1&type=pdf I think this is the classic summary paper of theirs, but they have published several others, including some more recent.