arXiv Limits Each Person to Two Paper Submissions a Month Starting in October
The rule covers every subject area. The count follows whoever actually uploads the paper, and a submission rejected in moderation still counts as one.
Starting October 1, arXiv, the preprint server where researchers post papers ahead of formal publication, limits each person to two new submissions per month. The window is a calendar month, and the rule applies equally across every subject area. Researchers in Korea fall under the same standard. The separate rule in place since 2024, which caps a person at three submissions being processed at the same time, also remains in force. That means researchers now have to track two things at once: how many papers they have submitted that month, and how many of their submissions are still being processed.
Why the cap dropped to two per month
Submissions have been piling up too fast. September submissions roughly doubled between 2024 and 2026, rising from 20,569 to 40,363. In September 2016, the figure was 9,869. Within computer science, submissions in the AI category have grown more than sixfold over the past two years. arXiv said that September’s submissions this year generated roughly 9,000 inquiries and support requests for staff and moderators.
arXiv pointed to AI tools as one driver of the surge. Papers with extremely narrow scope, single studies sliced into multiple submissions, and dense AI-written manuscripts have all increased, piling pressure on volunteer moderators. The organization also cited a problem with some authors dumping large batches of low-quality papers at once.
Limits on submission frequency existed before. In practice, though, enforcement was largely left to moderators’ discretion. Now the same two-per-month standard applies to every field. arXiv described the rule as a temporary measure while it overhauls its moderation process. It has not decided when the cap will end or when it might be loosened.
Does this count co-authors listed on a paper?
No. The count accrues only to the person who actually submits the paper. Posting to a different subject area does not reset or create a separate allowance. If you submit one paper to one field and another to a different field in the same month, you have used up both of that month’s slots. Submissions rejected during moderation also count as one. If you plan to resubmit after a rejection, it is worth checking how many slots you have left first.
The monthly count and the in-process count are calculated differently. A paper submitted last month that remains on hold into the next month does not count toward the new month’s two-submission limit. It does, however, continue to count toward the three-in-process limit. If all three papers you submitted last month are still being processed, you cannot post anything new even after the calendar turns. Deleting a submission before it goes public removes it from both counts.
What about AI research?
arXiv is not blocking research that uses AI. It does require authors to disclose that AI was used and to meet scholarly standards. The submission cap applies to AI-related fields just as it does everywhere else. If a group plans to post several papers, it is safer to settle in advance who will act as the submitter and to check how many papers that person has already submitted that month.
What to watch from here: how much arXiv’s changes to its moderation process ease the workload, and when the temporary limit gets adjusted as a result.