I suspect that they aren't factoring in the thermodynamic limits of the engines, which when applied will lead to a limitation on T FLEX.
There are further limitations that must be applied too.
I made a small app for calculating T FLEX based on the numbers from SimBrief, and even that results in sketchy take-offs now and then so there's more to it than that.
There is at least 1 bug in the TRP N1 calculation, which is that it uses indicated altitude instead of pressure altitude, which will result in different N1 REF values than the N1 values that all performance data tables are based on.
This could result in better and also worse take-off performance (I sent a ticket about this but nothing's happened yet).
If the QNH is low, the pressure altitude is higher and usually means a higher N1 REF is required.
Let's assume bleeds and eng ant-ice off, OAT 15 and the airport elevation is 0 ft.
Assuming we've set the correct QNH, that would yield an N1 REF of 94.2 in the JF RJ since it uses the indicated altitude, which would be around 0 under those conditions.
The pressure altitude would only be 0 at QNH 1013. But if the QNH is 978, the pressure altitude is around 1000 ft.
If we enter the N1 REF tables with 1000 ft instead of 0 ft, we get an N1 REF of 95.5.
Because the performance tables assume that an N1 REF of 95.5 is used under these conditions, SimBrief would calculate a higher MTOW than the JF RJ can actually manage since it will produce less thrust than IRL for those conditions.
On short runways, you may want to reduce the calculated max temperature by 10 degrees or something like that to give you greater margins and consider using N1 REF instead of N1 REDU or N1 FLEX.
I would've made a proper TKOF/LDG performance calculator long ago (completely free of charge), but no one is willing to give me the data I need, so sadly that will never happen.