The present study is concerned with the perception of the stop consonant clusters /pt/ and /kt/ and CVC-Sequences /pVt/ and /kVt/ in European and Brazilian Portuguese. A certain neutralisation of these lexical differences in both varieties has been attested (Bisol 1999, Mateus & d’Andrade 2000, Mateus et al. 2005, Vigário 2003), and therefore we tested the following three hypotheses: firstly, that listeners would not distinguish between clusters and CVC-sequences in Brazilian Portuguese (BP) and in European Portuguese (EP). Secondly, that they would be more likely to hear a cluster for the European and a CVC-sequence for the Brazilian data, thirdly, that the native listeners would be more likely to recognise subtle differences in the phonetic detail than non-native listeners. Twenty-one L1 EP and sixteen L1 BP speakers participated in a forced choice perception experiment in which they had to judge whether medial /pt/ and /kt/, /pVt/ and /kVt/, excised from real words that had been produced by one EP and one BP speaker, had been produced as a cluster or as CVC-sequence. The results showed support for the second, but not for the other two hypotheses. The general conclusion is that the lexical differentiation between stop clusters and CVC-sequences is neutralised in different ways, and that their perception depends more on the speaker’s variety and on the consonantal place of articulation than it does on the native or non-native listener’s varieties.