Thanks for the question. The first example in your documentation link shows how to download a model, as you've probably done. I do this by writing a python script and running it from the command prompt.
The script looks like this:
from allensdk.api.queries.biophysical_api import BiophysicalApi
bp = BiophysicalApi()
bp.cache_stimulus = True # change to False to not download the large stimulus NWB file
neuronal_model_id = 473862496 # here's your model
bp.cache_data(neuronal_model_id, working_directory='neuronal_model')
You can run this from the command prompt (Anaconda command prompt is fine) as follows:
$ python <your_script_name.py>
Moving down the documentation, the next step to running the model is to run the following on the command prompt:
$ cd neuronal_model
$ nrnivmodl ./modfiles # compile the model (only needs to be done once)
$ python -m allensdk.model.biophysical.runner manifest.json
First you step into the working directory you specified in the first script.
Next you run a NEURON binary (nrnivmodl), which compiles your modfiles. You'll need to have NEURON with python bindings installed and on your PATH to run this. I'm not sure about this, but I think compiling modfiles in Windows requires a different command/workflow. If that's your operating system, I'll have to refer you here since I'm not too familiar with NEURON on Windows:
https://www.neuron.yale.edu/neuron/static/docs/nmodl/mswin.html
Next you are calling an script packaged with the allensdk for running models based on one of the files we downloaded in the first script (manifest.json).