4 Kelvin is probably the correct number since that's the temperature of liquid helium.
4K is probably warmer than they'd like for a quantum circuit, but getting their apparatus colder would require a much beefier cryostat or more exotic technologies like magnetic cooling built inside a helium cryostat. Such setups are capable of getting complex setups down to the hundreds of millikelvin, which is much nicer for this sort of work!
As for writing a scientific article using Fahrenheit... Please don't do that. Just don't.
In the article they mention they chip itself is kept at millkelvin temperatures.
And I disagree about using Fahrenheit - this is not a paper in a journal, it's a lay article. By all means use Fahrenheit if that's what your audience understands. Don't let pedantry get in the way of clarity. (Kelvin has some real meaning, but C vs F is totally arbitrary. One is the freezing point of water, the other is the lowest temperature you can get with a salt/ice mixture.)