I think I know what you're saying, but comparing it with 051C is awkward, and making parallels between ships and aircraft is always dicey given the different nature of the key subsystems that make up their respective capability.
In the case of 051C, it was a ship that was only produced in a pair for a heap of reasons but perhaps first and foremost because it was a less risky air defence destroyer alternative if 052C faced major issues. As it was, 051C suffered from a variety of deficiencies ranging from its old steam turbine propulsion, using a less than ideal air defense system (foreign, lacking 360 degree PAR coverage), with other questionable design choices inherent to the pair (such no helicopter hangar as space taken up by the air defense system).
J-20 on the other hand, is an aircraft whose subsystems are virtually all or are soon to be all indigenously sourced. It is also not the "lower risk" alternative as 051C was to 052C, but rather J-20 is the only heavy weight stealth fighter the PLA will be buying in the near future.
Furthermore, any deficiencies that exist within the current J-20 variant will likely be corrected in future variants in an iterative manner in a far simpler way than it would've been to try to make a further variant of 051C.
Instead, I would say that the current J-20 variant (let's call it J-20A) would be more similar to 052C rather than 051C -- i.e.: not a "test bed" but rather the first design of a stealth fighter that they will produce in a decent (but not large) number that will provide key new capabilities in an operationally useful way, before moving onto a more "complete" variant (e.g.: having WS-15 and perhaps other iterative advancements) that will likely be produced in a larger number, like 052D.
So I agree that they probably won't produce many hundreds of the current J-20A variant, but I also think that they will produce enough such that they won't merely be "test aircraft" but rather form the core of an operationally useful stealth fighter capability, to be followed up with by a more complete J-20 variant with WS-15 engines and other improvements that will be mass produced in the longer term. Of course, if WS-15 is significantly delayed I imagine J-20A production using interim engines (Al31 or WS-10 variant) to continue for longer in which case more J-20As may be produced than otherwise expected.
If I had to compare the current J-20A to a PLA ship class, I would say it is best compared to 052C, with a follow on "J-20B" (with WS-15s etc) likely to be the equivalent of 052D -- but even such a comparison is not perfect.
