Running strings on the provided mp3 resulted in a lot of layer/... strings.
Asking chatgpt if this is normal behaviour for a mp3 file it returned that it looks like a docx file.
Running file on the mp3 confirmed that the mp3 file is a zip.
I extracted the contents with binwalk.
The flag format of the ctf is grey{<flag>}.
I ran `rg grey` in the directory of the extracted files and was presented with the flag.