arXiv, the free site where scientists post papers before peer review, said on October 1 that each submitter may now send at most two papers per calendar month, in every subject. The rule started the same day. arXiv says AI tools are helping a small group of authors flood its volunteer moderators with low-quality work.
New to this? Read it in simple words
- arXiv is a free website where scientists share new papers.
- Now each person can send only two papers a month.
- It got a record 40,363 papers in September.
- arXiv says AI tools help some people send many weak papers.
- Preprint
- A science paper shared before other experts have reviewed it.
- Moderator
- A person who checks posts before they appear on a site.
- Submitter
- The person who uploads the paper, even if there are many authors.
The new rule
From October 1, a submitter can send at most two papers per calendar month and have at most three waiting at any time, arXiv’s blog says. Before, limits were mostly left to moderators’ judgement.
Rejected papers count toward the two. Papers deleted before they are announced do not, arXiv says. Co-authors do not use up any of their own quota.
From arXiv’s blog post of October 1, 2026.
The numbers behind it
arXiv received 40,363 submissions in September 2026, a record. In September 2024 it received 20,569, and in September 2016, 9,869.
“AI tools are making it easy for authors to flood arXiv and other repositories with these low-value papers,” the blog post says.
“A relatively small proportion of authors are submitting a large number of low-quality papers and consuming a disproportionate fraction of the moderators’ time,” it adds. These are arXiv’s own figures and reasons.
How researchers reacted
Times Higher Education reports a split. Mathematician Alex Kontorovich backed the cap: “Pick two, your very best two.”
Economist Scott Duke Kominers said arXiv should “economise on moderation – not on scholarship”, according to the same report.
Mathematician Terence Tao wrote about the change on his blog the day it was announced.
Sources
Every fact in this story comes from the sources below. Open them to check our work.
- 1Primary source · October 1, 2026Fair Moderation, Equitable Access, and AI: arXiv’s Updated Rate Limit Policy arXiv
- 2Research · October 2, 2026arXiv’s new preprint submission cap divides opinion Times Higher Education
- 3Research · October 2, 2026arXiv now limits every researcher to two paper submissions a month Startup Fortune
- 4
We read arXiv’s blog post and compared Times Higher Education, Startup Fortune and Terence Tao’s blog. The submission counts are arXiv’s; we could not read them from its statistics page ourselves. Some headlines say the cap is per researcher or per author; arXiv’s rule counts only the person who submits, so we use that.