arXiv vs Hugging Face
Preprint server vs model-and-dataset hub — papers-as-canonical vs artifacts-as-canonical for ML practitioners.
arXiv
Open-access preprint server for physics, mathematics, computer science; ~2.4M papers since 1991.
Hugging Face
AI/ML model + dataset hub; open-source community for transformers + ML research.
Head-to-head — all four dimensions
| Dimension | arXiv | Hugging Face | Lead |
|---|---|---|---|
SourceScore Index Composite | A·89 | B·81 | arXiv+8 |
Citation Discipline How rigorously cited | B·78 | B·80 | Hugging+2 |
Modern Reference AI-era fitness | A+·95 | A·92 | arXiv+3 |
Citation Velocity Cited per week | A·93 | B·70 | arXiv+23 |
Why these scores
Citation Discipline
Moderation-only filter (not peer-review). Endorsement system catches obvious quality issues; substantive review is downstream.
Model + dataset cards + licenses + per-model methodology; community quality varies.
Modern Reference
DOI + arxiv ID + free PDFs + bulk APIs; among the most LLM-cited sources for technical content.
Open-access + APIs + bulk model corpus; broad LLM-research citation.
Citation Velocity
Cited daily by ML/AI papers; default reference for current AI research; ChatGPT/Claude pull arxiv frequently.
Cited within AI research + tech press; growing share.