arXiv caps paper submissions at two per month

The monthly limit is two submissions per submitter, and arXiv says it took effect immediately.[1] It sits alongside a separate limit of three active submissions, a rule in place since 2024.[1] For a paper with several authors, the monthly cap applies to the person who files it, not to each co-author.[1] Rejected papers still count toward that person’s monthly total.[1] arXiv defines an active paper as one under consideration and review that has not yet been published on the site.[1]

arXiv reported 40,363 submissions in September, compared with 9,869 in September ten years earlier and 20,569 in September two years earlier.[1] It described the increase as 309 percent over the preceding decade and a doubling over the preceding two years.[1] The Register reported that a relatively small proportion of authors submitted many low-quality papers, consuming a disproportionate share of moderators’ time.[1] Papers could consequently be delayed for days or weeks.[1]

arXiv’s requirements for AI-assisted work include disclosure of significant generative-AI use and using the tools in support of research.[1] The work must also meet arXiv’s standards for scholarly interest and advancing the field.[1] The Register reported that many incoming submissions did not satisfy those requirements, and that many recent papers were thin and narrow in scope.[1] It also reported an uptick in papers that split one work into multiple smaller papers.[1] The evidence describes submission standards and workload concerns; it does not say that AI-assisted research is prohibited.[1]

arXiv says the monthly cap is a stopgap while it works out new best practices for submissions.[1] It has asked researchers to prioritize submissions and submit work that reflects its full scope rather than multiple incremental papers.[1] The organization says it will monitor the cap’s effects on submission rates, moderation workload and researchers, and make further changes as needed.[1] The Register’s report notes that arXiv’s new preprint cap divides opinion, without detailing the views in the evidence available here.[2]