Apologies if you're already aware of this but the 'correct' process is to tweak the powder burning parameters til the model matches your MV. With these internal ballistic models the old maxim "garbage in = garbage out" applies absolutely.
There are many variables in your set-up that may be slightly different from standard - barrel diameter could be slightly different, powder scale could be marginally out, chrony accuracy could be slightly out, measurement of internal volume of cases could be out etc etc. Hence trying to use accurate as possible data points and minimising the variables you adjust. Details are important.
One thing I would say with confidence, the reason GRT's estimate and your measured data are different is NOT because their algorithm has an error with muzzle length. (And I know you're not suggesting this)
By the way, version 2021.1980 is not the latest version, although I doubt the version makes much difference but there might be more refined powder data in the latest version (.2030)
Just because something works by coincidence doesn't make it good practice - I don't think you'll have issues modelling with a shorter barrel from a safety perspective (fudging by using a longer barrel than your actual length I would worry about). The powder burn calc and OBT will be totally out of the window because the barrel travel/residence time is very dependent on precise barrel length. This of course may be unimportant to you.
One of the great strengths of GRT is that users can easily submit their empirical data back to GRT and they can use the data to fine-tune their powder algorithms, which all users benefit from.