First, create a file named gurobi.env with these contents:
Save that file to the current working directory for your program that uses Gurobi. That is usually the directory from which you launch the program. That might be a compiled application that links to Gurobi or it might be a run-time environment like Python.
After saving the above file, run your program or model. Then look in the directory where it ran for the recording file generated by Gurobi named, recording000.grbr. If there is more than one recording, the file name is incremented for each new one recording001.grbr, recording002.grbr, etc.
The file is a step-by-step "recording" of the Gurobi API's used in your code. From this recording we can extract the model and test the actual API's you used to build it.
This file should be uploaded to the Gurobi support team. Instructions for uploading files can be found here.