I agree with your current approach. Obviously we don't have much formal guidance of foreign body size (and they're arguably inherent rather than foreign in origin ) but larger pips along the longest dimension could perhaps challenge the 7mm FDA limit that's often a handy guide.
If you're defining intended user as all age groups then this should definitely also be considered - a quick google suggests nuts and seeds are a very common choking hazard for young children.
In terms of quantifying risk, the likelihood of presence is also presumably fairly high, unless you have an unusual extraction process and/or other controls for this prior to the final sieve?
I've just had a quick look at process flow diagrams / CCP details for all of our citrus juice suppliers (spanning small to larger European and South African operations and extremely large Brazilian sites) to check my own thought process and these all set the final sieve as a CCP, but to be honest that's fairly typical for juice processors in general IME, even for products without such an obvious potential for pips.
If nothing else it means that you know you are in control of it, and I really can't see any regulator or auditor quibbling with it either.