openai/gpt-5.2-chat → A (2:1)

Side A introduces a more substantial behavioral change by redesigning the rank row coloring logic from ordinal-based to score-based min–max normalization, modifying function signatures, updating call sites, and adding multiple focused tests to validate edge cases and invariants. This improves visual semantics and scalability of the UI. Side B adds a useful but relatively small filtering enhancement (skipping pinned posts) with a helper and a single test. Overall, Side A represents a broader and more impactful contribution.