News · 2026-10-02

arXiv’s published policy caps authors at two monthly submissions

arXiv’s official moderation policy limits each author to two submissions per calendar month and three active submissions. The numerical rule changes how authors can send work into the preprint platform, while the accessible primary documentation leaves several collaboration and quota-accounting details unresolved.

Key facts

• The verified limits are two monthly submissions and three active submissions.
• The rule appears in arXiv’s official moderation policy across categories.
• An indexed announcement copy dates the change to October 1; that date was not confirmed on the inaccessible original announcement.
• The primary source is arXiv’s submission-rate policy.

arXiv is a major route by which researchers share results before journal publication. For AI readers, it is also a daily source of claims, methods, datasets, and benchmarks. A restriction on intake therefore affects the research information pipeline even though it says nothing directly about the validity of a particular paper.

The policy’s exact wording is clear: “Each author may submit up to two submissions per calendar month with a limit of three active submissions.” That provides two separate constraints. One controls flow over a month. The other controls how many submissions can remain active at once. Neither is the same as a limit on how many papers a person may write.

Think of an office that allows two new applications per month and no more than three unresolved applications in the queue. The monthly limit controls arrivals. The active limit controls outstanding work. An author with unresolved submissions may need to consider both. The analogy explains the distinction, but it should not be stretched into an interpretation of undocumented arXiv counting rules.

The moderation overview says moderation happens primarily between submission and public announcement. It also makes clear that moderation is not peer review. That distinction is essential when reading both the policy and papers hosted on the service: acceptance into a subject category is not a journal’s judgment that the results are correct.

The broader announcement is less fully verified in this dossier. An indexed reproduction describes the measure as a stopgap while arXiv improves moderation and considers practices around advanced AI tools. It also reports substantial submission and support volume. The original announcement could not be opened, so those figures and its stated motivation should not be promoted into primary-verified facts in this article.

The confirmed story is consequently the policy itself. It is fair to discuss the possible operational purpose of a rate limit, but a reader should distinguish that explanation from a documented causal finding. A cap can reduce how much material reaches an intake queue under the stated rule. It cannot, by itself, distinguish good science from bad science or identify text written with AI.

Several practical questions remain open. The primary material checked does not establish that the quota belongs only to the uploader’s account, or explain how every co-author’s quota is counted. It does not verify a collaboration exemption, treatment of deleted or rejected submissions, or how cross-listing affects the count. The more specific claim that all those cases have settled answers would exceed the evidence.

The replacement instructions tell authors to replace corrected articles rather than submit them again as new work. That is useful procedural guidance. It does not explicitly say how replacements interact with the monthly limit. The submission overview provides broader process information, but cannot be used to invent a quota exception.

The reception evidence is modest. The accessible MachineLearning discussion contains the news post and one visible approving comment in the dossier’s check. The intended Hacker News thread was inaccessible. That record does not support claims of researcher consensus, widespread outrage, or a representative endorsement.

The strongest policy objection is straightforward: a uniform volume cap can constrain legitimate prolific researchers and collaborations as well as low-value submissions. The strongest operational argument is that finite moderation capacity needs some workable intake control. These are policy trade-offs, not measured outcomes of the new rule. Whether the balance is good will depend on implementation, exceptions, and effects that the present evidence does not establish.

For daily AI coverage, the change reinforces a separate reading discipline. A preprint is a public research claim to inspect, not automatic validation. Our lessons on benchmark contamination and preregistration explain two ways to evaluate evidence beyond publication venue. arXiv’s confirmed rule is significant infrastructure news; its exact motivation, edge-case behavior, and eventual consequences still require documentation rather than inference.

Primary source, verified: read the paper →

Ground Truth. (2026, October 2). arXiv’s published policy caps authors at two monthly submissions. Ground Truth. https://groundtruth.day/news/arxiv-two-submissions-three-active-policy.html

@misc{groundtruth:arxiv-two-submissions-three-active-policy,
  title  = {arXiv’s published policy caps authors at two monthly submissions},
  author = {{Ground Truth}},
  year   = {2026},
  month  = {oct},
  url    = {https://groundtruth.day/news/arxiv-two-submissions-three-active-policy.html}
}

Topics: policy · research · publishing · scientific-infrastructure