Getting parts space-qualified is a very expensive and lengthy process. It does not make any economic sense to build one-offs.
In particular, unlike any other currently operating orbital launch vehicle, which just lets its payload fairings burn up in the atmosphere, SpaceX recovers and re-uses its payload fairings. You cannot reuse a payload fairing that was built specifically for one launch.
To the best of our knowledge, a pair of fairing halves costs about $5 million. (Note: this number comes from a tweet by Elon Musk where it is not clear whether he means one fairing half or an entire fairing consisting of two halves.) That number, however, is predicated on these fairing halves being standardized, such that R&D costs, qualification costs, testing costs, one-time setup costs for machines, etc. can be amortized across the manufacture of many identical fairings … or to put it differently: the laws of economics don't just magically change just because the thing you are building goes to space.
Building a one-off fairing for DART would very likely cost significantly more than $5 million and would not be able to be amortized across many re-uses. I can only think of one other comparably small payload in recent times, which is the IXPE spacecraft. This is the only payload right now I can think of which could maybe have reused that fairing … except it launched only two weeks after DART, so it would probably not have been possible to refurbish the fairing in time.
There is another aspect to consider: launch cadence. Before SpaceX started reusing fairings, the manufacturing time for the fairings was actually the limiting factor for their launch cadence. They could not launch any more often because they couldn't build the fairings fast enough. And that was a couple of years ago, when they were launching 30 times a year, not 100 as indicated by Elon Musk for 2023.
Also note that SpaceX is currently developing a differently-sized fairing, although it is a bigger one rather than a smaller one. This development has been ongoing for about two years now.
Let's look at the two assumptions in your question:
It reduces mass and is definitely cheaper
As I explained above, I am not convinced the second is actually true. Mass production is generally cheaper per unit than one-off production.
The first one is obviously true, but the question is: does it matter? Remember, you would only use a smaller payload fairing for a smaller payload … but smaller payloads are already lighter than others! There is not much reason to reduce the mass if you already have enough performance.
See the IXPE launch for example. Inclination changes are ridiculously expensive, and yet, SpaceX was able to pull off an inclination change of almost 30° (just look at the trajectory projection on the map) and still recover the first stage. The only improvements they could have made here would be to recover the first stage on land instead of at sea, but shaving off a couple of kilograms from the fairing wouldn't have made a difference here.