Accepted answer
This is the question where methodology matters more than the conclusion.
Submitting to different laboratories introduces a method difference that commonly exceeds the supplier difference. Gradient slope alone can move a reported purity figure by a per cent in either direction.
Compare documentation on a fixed checklist rather than by impression: lot specificity, method section, chromatogram availability, quantified content, water and counter-ion figures, and whether the code is on the vial.
Content assay results across the published datasets vary considerably more between suppliers than purity does, which makes content the more discriminating axis.
Compare content, not purity. Purity clusters and content does not.