Moved from comments.
You might consider generating such lists yourself. I'm not saying asking if that work has already been done is wrong. :) Computationally, it's not a trivial job. This link (billposer.org/Software/minpair.html) purports to provide an open source (GNU license) program to find minimal pairs in a provided dictionary. You'd want to provide a phonemic dictionary, of course, for phonological minimal pairs. If it's not adequate, you could also examine the algorithm and modify it to suit your needs.
You might also be interested in generating your own words (according to some pattern of phonemes, presumably), and then simply checking if they are in the dictionary. Or, and this is based on your original question, you can test people on minimal pairs of nonwords (in something like an AXB task, I guess); for the purpose of assessing their phoneme perception, this option is the typical choice. It helps to minimize the influences of semantics, frequency, lexical neighborhoods, etc.
@JeremyNeedle: I like using real words. Given a phonemic dictionary I have no issue with doing the programming work. Where do I find the computer readable phonemic dictionaries? – Christian
That can be a little trickier, and it's language specific. For French, I believe Lexique is a good choice (http://www.lexique.org).
Researchers have also had success using eSpeak (http://espeak.sourceforge.net) to convert orthography to phonemes, for French and other languages (for eSpeak performance figures, see Marian, V., Bartolotti, J., Chabal, S., Shook, A. (2012). CLEARPOND: Cross-Linguistic Easy-Access Resource for Phonological and Orthographic Neighborhood Densities. PLoS ONE 7(8): e43230. doi:10.1371/journal.pone.0043230).