@samarth jain: The issue does not exist in the open source PDK because I fixed it. The continuous-binned models fail to cover the short device used in some of the standard cells. I had to make a copy of the single bin for that device from the discrete-binned models and then create a new device type for it. Apparently Cadence is way behind me.

i see,so u made a copy and exactly used same device length and width, just it was issues with modeling? I was scared if the pdk was having wrong device width

I think the underlying issue is that the people who did the continuous-binned models did not include the device length from the standard cells. That specific device is specified in the DRC rules as being shorter than normal, and the DRC rules allow for it to exist in the standard cell library only.