> the puppy's calculation on KC is coming out at 9.6%
That makes sense as you are combining the same grandparents in this generation, but I wouldn't get hung up on numbers, and I wouldn't call 9.6% high at all.
Speak to the breeder, ask about the line breeding, what do they want to solidify and are there any potential problems being doubled up on. It's a perfectly reasonable breeding, not too close, providing the breeder knows the dogs they're doubling up on well, or knows someone who does.
As has already been said the difference between your calculation and the KC's is purely about the number of generations used. In my pedigree programme I set it to use every available generation, which gives a more accurate, albeit higher, number.
To give you a benchmark, 30 years ago most show dogs had COIs in the 20s or 30s, increased awareness, cheap programs that work it out and the pet passport and cheaper, easier imports have brought COIs down significantly in most breeds.