d
Franz
I apologize but this is kind of useless quibble until as Franz said, they are both run on the same server.
Logically, you cannot run a fair benchmark until they are even. However, you can run a uneven benchmark
that favors one or the other to make a point.
I have a 512MB 2 core VPS. Logically, I should be at a disadvantage with half the needed ram. I'll be comparing it to the primary discourse try it site (which I assume has 1gb+ of ram)
To make it more even, I am clicking on a discussion with 1 post and no added graphics. Then I mimic this for my demo flarum (so it has similar content - which I actually copy multiple times)
There are a lot of reasons why this is not a valid test, so make your own judgements.
Summary:
~400 response time discourse
~180 response time flarum

Edit: Full disclosure - I run the Beta Breaker flarum site tested above and I have personally optimized it. Originally it was getting like 260ms or so and now I (as seen) get about 180ms. I think if you try this site itself, it will load at about 260ms. This kind of brings us full circle back to it's not just about the forum software, but a great amount is webserver config too.