Yeah I get the same error. Wrote up my findings on this thread: 500 error in request to gpt-4o-audio-* model - #13 by pj4533