what did you mean that the band gap is incorrect? Although the DFT method often under estimates the band gap of semiconductor, it can be a good approximation. There were many different values in publications using DFT method, maybe from 1.5eV to 1.9eV. Despites the deviation of the gap value, the characteristics of the band structure are similar, that is, direct band structure at point K. If you do think that DFT-GGA/LDA is not so good, you can try DFT-MGGA.
