There's no hard and fast rule, since it depends on how hard your test server will be working relative to the prod server. Ideally they should have the same configuration so you can get consistent behavior when you're testing, but if you'd like to save some resources:
Disk - no difference, really. You'll have the same apps and files on both servers.
CPU - if the PROD server is handling many scheduled reloads and the Test server doesn't have to, you might be able to save some CPU, but you only have two processors in the first place so I'd just stick with that.
RAM - If the test server is only likely to be handling a handful of apps / people at a time compared to the prod server, you could probably cut this down. This does also depend on the size of your individual apps. Most typically, between 1/2 and 1/4 of the PROD server might work but again this depends on assorted factors and is just a rule of thumb. If you don't assign enough RAM, this won't generally break anything, but it will cause much slower performance.
As @Or is saying there are no real rules here. But the closer the specs are, the more "realistic" the tests will be, especially in regard to performance. But again it depends on how you are planning to use it.
But the memory is very important and it can bring the whole test server (Windows Server) down if you do not assign enough.