In which kind are the values different? The values itself or the number of rows? Often it's helpful to compare directly certain records - maybe you have an ID or you creates a row-number. Within the load per rowno() and in select per https://599cd.com/tips/access/140703-row-number/.
I would suggest you make a copy of the query and simplify it, remove aggregations, so you see the underlying records, and see if you can spot which rows are different.
Things I would look out for are for example dates or decimal numbers stored in text fields, which could be interpreted differently if the query is run with different regional settings in the two different programs, or maybe null values.