This paper describes the development op- erated into MANY, an open source sys- tem combination software based on con- fusion networks developed at LIUM. The hypotheses from Chinese-English MT sys- tems were combined with a new version of the software. MANY has been updated in order to use word confidence score and to boost n-grams occurring in input hypothe- ses. In this paper we propose either to use an adapted language model or adding some additional features in the decoder to boost certain n-grams probabilities. Ex- perimental results show that the updates yielded significant improvements in terms of BLEU score.