Pi wrote:It would be great if someone could brute force it to check that it is not valid without the inequalities being taken into consideration
You could keep just the one inequality between r9c4 & r9c5, and delete all the rest -- the puzzle would still have only one solution, and be very easy to solve. (But deleting them all
would produce a puzzle with two solutions.)
Personally, I prefer the type of inequality puzzle in which all the digits are absent, and there are inequalities between adjacent cells within each box. Every completed number-grid converts of course to exactly one (possibly invalid) inequality puzzle of that type.
It might be interesting to try to make a variant of that kind, going one step further and deleting as many inequalities as possible -- with the intention of having a unique-solution puzzle with only blanks and a minimum number of inequalities