Strangeness production

I would like to compare different event generators (UrQMD, DCMQGSM-SMM etc.) with GEANT4. Particular: strangeness production. I used QGSP_BERT_EMV physics list to simulate Au beam @ 8A GeV/c + Au target. I got ~5 times less number of kaons and ~10 times less number of Lambda and K-short with GEANT4, than with event generators. I tried FTFP_ INCLXX, which, as I read, has improved strangeness physics. But I didn’t see any difference in number of strange particles. Are they expected results or are there some problems with my simulations? I use Virtual Monte Carlo.