DRY isn't always what you want with tests - the more important factor is legibility so its easy to understand what is under test, what the expected values are, and how you might write similar code to exercise the same behavior in the source code.

I find this much harder to read - I think this either needs quite a bit more comments to explain what is under test or needs to be refactored.

As an example, reading a line like:

it('Contains values', check(type, [1, 2], expected.containsValues));

I have no idea what is being tested, how [1, 2] relates to a tested or expected value, or what expected.containsValues are.

Similarly:

allChecks(GraphQLList(GraphQLInt), {
    containsValues: { data: dataOk },
    containsNull: { data: dataOkWithNull },
    ...
})

I'm not sure what this is testing, or what the expected value is.