The artifact contains the replication package including code, benchmark models, datasets, and results for the ICSE 2023 paper entitled “Fairify: Fairness Verification of Neural Networks”. Fairify is an SMT based approach to verify ReLU based neural networks (NN). We created a benchmark of NN models trained on 3 datasets, e.g., Bank Marketing (BM), Adult Census (AC), and German Credit (GC). Fairify is implemented in Python and openly available SMT solver Z3. Input. Fairify takes the trained NN model, input domain, the verification query, maximum partition size, and timeout as inputs. The trained models are given as h5 files. Output. Fairify provides verification result for each partition. The results include verification (SAT/UNSAT/UNKNOWN), counterexample (if SAT), and pruned NN for the partitions. Running Fairify is automated through terminal commands and results are saved into CSV files. The usage instructions and further details can be found in our GitHub repository: https://github.com/sumonbis/Fairify.