Some of your devices are inaccurate. The longest thing I ever recorded using multiple devices was a Matthew passion, almost 3 hours long in one run. The cameras we used were pro or prosumer cameras, and there was no noticeable audio drift.
However cheap devices can have varying speed. At the moment I would blame your audio recorder, but both device may have very small inaccuracy which at the end sum up.
Can you do some dry-run tests in a silent room for as long as your usual flights?
Just clap around every 10 minutes.
Then sync that as usual, and see if the drift is constant?
If it is not constant I'm afraid there's not much help.
If it is constant, so the first claps are in sync, the second differs 10ms, the 3rd differs 20ms, 4th differs 30ms, etc, then you can do this:
Sync the first claps.
Then sync the last clap with ctrl-drag, so that length of your audio will change.
I'd expect this a very low percent change, say 99.997% or 100.002%. So you could ignore the pitch change, and just allow Vegas to pitch change locked to stretch.
Just one more thing, before you start the procedure above:
I'm not sure what option you have on your recorder/gopro, but if the audio sampling rate don't match, try to set them to match. So both to 48000, or both to 44100.
Avoid having one set to 48000 and the other to 44100.
If you can't do that, you can still try what I already suggested. ;)