These are the differences between CPython (the normal Python implementation) and Nuitka:
Is not bytecode. There is no bytecode anymore with compiled functions.
threading can block it seems
The generated code never lets the CPython run time switch threads, so its chances to do so are reduced, which may lead to problems. I personally do not care much about threads, would use subprocesses anyway. So for the time being consider threads unsupported.
Start of function call vs. end of function call in traceback output
In CPython the traceback points to the end of the function call, whereas Nuitka has it point to the first line of the function call. This is due to the use of the ast.parse over bytecode it seems and not easy to overcome. It would require parsing the source on our own and search for the end of the function call.
It's a really small amount of things, isn't it? I consider that there are practically no differences therefore.