We study the data on mean hadron yields and contrast the chemical freezeout conditions in p+p, p+Pb and Pb+Pb at the Large Hadron Collider (LHC) energies. We study several schemes for freezeout that mainly differ in the way strangeness is treated: i. strangeness freezes out along with the non-strange hadrons in complete equilibrium (1CFO), ii. strangeness freezes out along with non-strange hadrons with an additional parameter $gamma_S$ accounting for non-equilibrium production of strangeness (1CFO+$gamma_S$), and iii. strangeness freezes out earlier than non-strange hadrons and in thermal equilibrium (2CFO). A comparison of the chisquares of the fits indicate a dependence of the freezeout scheme on the system size. The minimum bias p+p and different centralities of p+Pb and peripheral Pb+Pb data prefer 1CFO$+gamma_S$ with $gamma_S$ approaching unity as we go from p+p to central p+Pb and peripheral Pb+Pb. On the other hand, the mid-central to central Pb+Pb data prefer 2CFO over 1CFO+$gamma_S$. Such system size dependence of freezeout scheme could be an indication of the additional interaction in Pb+Pb over p+Pb and p+p.