You should not use M to define a multi-finger device.
Most non-linear device model calculations are based on the active area of the semiconductor device. Some devices do break this down to width, length and number of fingers and provide parameters to set each of these values but most only give the user the area parameter itself and it is up to the user to decide how this is to be calculated. The area is also used to calculate parasitics that are dependant on the area and the perimiter of the device. Using M to scale the area of the device would scale the area but not the perimiter parasitics correctly. It just gives you M separate devices connected in parallel and not a single device M times larger.
As erild suggests Foundry PDKs usually implement their devices in a way that users can set the physical size of the device and the number of fingers. They then internally calculate the necessary device size depending on the process characteristics. Every foundry does this their own way so you need to understand the process itself if you are to do this scaling correctly.
With a PDK everything is defined for you, without a PDK you would need much more detail of the manufacturing process to continue. Especially if you intend to eventually create the layout for the circuit and have it manufactured.