Comparing NFT rarity meters and benchmarks

Non-fungible tokens (NFTs) have established themselves as a new class of digital assets with unique ownership properties verified through blockchain technology. Their application in digital art has given rise to NFT collections where each token, generated by smart contracts, combines distinctive visual characteristics with embedded metadata. Rarity as a critical value determinant is examined through modern evaluation methods, encompassing platform-native solutions (OpenSea, Rarity.tools) and scientific methodologies. This work presents a comparative analysis of six prominent rarity metrics (Rarity.tools, OpenRarity, GoRarity, etc.) using ROAR and DIT benchmarks on data from 100 major collections. The experimental results reveal significant differences in computational complexity and market price correlations, demonstrating the superiority of hybrid approaches that synergize trait analysis with historical transaction-based optimization.