Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces a four-tiered taxonomy of 646 programming languages based on their resource availability for NLP and code generation tasks. The authors quantify the extreme resource imbalance across languages, demonstrating that a tiny fraction (1.9%) of "High" resource languages dominate code corpora. This taxonomy enables more principled dataset curation and evaluation of multilingual LLMs for code.
Just 2% of programming languages account for nearly 75% of code tokens, revealing a winner-take-all dynamic that should inform how we train and evaluate code-generating LLMs.
The world's 7,000+ languages vary widely in the availability of resources for NLP, motivating efforts to systematically categorize them by their degree of resourcefulness (Joshi et al., 2020). A similar disparity exists among programming languages (PLs); however, no resource-tier taxonomy has been established for code. As large language models (LLMs) grow increasingly capable of generating code, such a taxonomy becomes essential. To fill this gap, we present the first reproducible PL resource classification, grouping 646 languages into four tiers. We show that only 1.9% of languages (Tier 3, High) account for 74.6% of all tokens in seven major corpora, while 71.7% of languages (Tier 0, Scarce) contribute just 1.0%. Statistical analyses of within-tier inequality, dispersion, and distributional skew confirm that this imbalance is both extreme and systematic. Our results provide a principled framework for dataset curation and tier-aware evaluation of multilingual LLMs.