Array capacitors - how well matched are they?

[axemaster]

I'm building a circuit that requires quad well-matched capacitors, NP0, in the 50-500pF range. "Well-matched" means ideally within 2%. Each cap will be from a signal line to GND, so bussed caps are OK.

I've looked at X2Y caps but they only provide 2 matched capacitors.

I also looked at standard array capacitors, which do have 4-cap arrays. But, the datasheets don't say anything about the matching.
Example digikey page: https://www.digikey.com/short/92m3q85q

I will be using several thousand quad caps, so manual measuring and matching is not a viable option here...

Does anyone have wisdom to share about this?

[TimFox]

The Kemet data sheet for the C0G quads says nothing about matching.  The only advantages they quote over singles is volumetric:  i.e., they take up less board space.
Obviously, they are not intended for matching tolerance better than the absolute tolerance, as opposed to some resistor networks.

[wraper]

I've looked at X2Y caps but they only provide 2 matched capacitors.

They don't specify matching tolerance at all. And according to my measurements of higher capacitance types (X7R, >=1nF) it's not better than taking 2 random capacitors from the same batch. IMHO just use separate capacitors with 1% tolerance, they are not that much more expensive than 5%, and most likely will be cheaper than a 5% array anyway.