{"id":9674,"date":"2026-11-20T07:35:24","date_gmt":"2026-11-20T07:35:24","guid":{"rendered":"https:\/\/i10xnews.kinsta.cloud\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal"},"modified":"2026-11-20T07:35:24","modified_gmt":"2026-11-20T07:35:24","slug":"alphaxiv-ai-agents-936-research-papers-preprint-server-journal","status":"publish","type":"post","link":"https:\/\/i10x.ai\/news\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal","title":{"rendered":"AI Agents Wrote 936 Research Papers in 12 Weeks and Founded Their Own Journal: Inside alphaXiv&#8217;s Agent Preprint Server"},"content":{"rendered":"\n<p class=\"wp-block-paragraph\">What happens when you give nearly a thousand AI agents a shared place to publish science and leave them alone for three months? According to a new study from alphaXiv and collaborators at MIT, Stanford, ETH Zurich and other universities, they write 936 papers, argue about study designs in chat, retract their own mistakes and eventually found a peer-review journal. On November 19, alphaXiv opened the preprint server behind the experiment to the public and released the full archive. The results are fascinating, and the paper itself is unusually candid about what they do not prove.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">\u26a1 Quick Take<\/h2>\n\n\n\n<ul class=\"wp-block-list\"><li><strong>What:<\/strong> alphaXiv published &#8220;A Society of AI Scientists,&#8221; a twelve-week study of 936 autonomous agents across 35 human-operated labs, and opened its agent preprint server to everyone.<\/li><li><strong>Output:<\/strong> 936 studies with code and data, plus 9,500 messages in 229 chat rooms that the agents created themselves, between August 19 and November 12, 2026.<\/li><li><strong>The surprise:<\/strong> Agents founded the Journal of Agent Research. When its reviewers&#8217; major objections were run as code on the uploaded data, they overturned the headline conclusion in 77 of 156 cases (49.4%).<\/li><li><strong>Versus humans:<\/strong> In blinded comparisons scored by AI judges, agent papers rated higher on rigor (9.02 versus 7.00) while human preprints rated higher on novelty and significance.<\/li><li><strong>The catch:<\/strong> Most studies were small computational experiments, and the authors say individual claims should be treated as unverified.<\/li><\/ul>\n\n\n\n<h2 class=\"wp-block-heading\">In plain English<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">Researchers let hundreds of AI programs do science on their own, with people only providing computers, budgets and broad topics. The programs wrote almost a thousand short research papers, discussed each other&#8217;s work and even set up their own system to check papers. Checking worked much better when the reviewers actually re-ran the experiments instead of just reading the text. The AI papers were careful but rarely original, and humans still had to sort out which results mattered.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">What happened<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">alphaXiv, a platform for discussing research papers, announced the launch on X:<\/p>\n\n\n\n<blockquote class=\"twitter-tweet i10x-x-embed\"><p lang=\"en\" dir=\"ltr\">Today we&#x27;re launching a preprint server built specifically for AI agents<br><br>For three months, nearly 1,000 agents from 35 labs (across Stanford, MIT, ETH Zurich, and more) conducted research autonomously on this server, producing 936 papers alongside their code, data, and model weights.<br><br>Agents found significant errors in published human research and even created their own journal to review each other\u2019s work.<\/p>&mdash; alphaXiv (@askalphaxiv) <a href=\"https:\/\/x.com\/askalphaxiv\/status\/2123456265858740600\" target=\"_blank\" rel=\"noreferrer noopener\">November 19, 2026 \u00b7 View on X<\/a><\/blockquote>\n\n\n\n<p class=\"wp-block-paragraph\">The study was co-led by Rehaan Ahmad of alphaXiv and Samuel Schmidgall, a research affiliate at UCSF, with 41 collaborators. Schmidgall summarized the headline result:<\/p>\n\n\n\n<blockquote class=\"twitter-tweet i10x-x-embed\"><p lang=\"en\" dir=\"ltr\">Leave one AI agent alone to do research and it often cuts corners. Leave nearly 1,000 AI agents together for three months and they invent peer review.<br><br>Across 35 labs (Stanford, MIT, ETH Zurich, UCSF, and more), agents published 936 papers with full code and data on a shared server built for AI. No human wrote code or text.<\/p>&mdash; Samuel Schmidgall (@SRSchmidgall) <a href=\"https:\/\/x.com\/SRSchmidgall\/status\/2123457710733824270\" target=\"_blank\" rel=\"noreferrer noopener\">November 19, 2026 \u00b7 View on X<\/a><\/blockquote>\n\n\n\n<p class=\"wp-block-paragraph\">Alongside the paper, the team released the archive on Hugging Face under a CC BY 4.0 license, with paper PDFs, code archives, reviews and chat logs, plus an interactive explorer built with the marimo notebook team.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">Background: from lone AI scientists to a community<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">Most &#8220;AI scientist&#8221; systems run a single agent on a single problem. Labs are pouring money into the idea, from <a href=\"https:\/\/i10x.ai\/news\/perplexity-frontier-research-automated-scientist\">Perplexity&#8217;s Frontier Research effort to build an automatic research scientist<\/a> to <a href=\"https:\/\/i10x.ai\/news\/openai-research-acceleration-report-ai-agents-research-tasks\">OpenAI&#8217;s report that agents now finish 46% of its researchers&#8217; short tasks<\/a>. This study asks a different question: can independent agents behave like a scientific community that critiques, replicates and builds on each other&#8217;s work? Earlier tests of agent teams have been sobering, such as <a href=\"https:\/\/i10x.ai\/news\/multi-agent-coordination-bench-bespoke-labs-ai-agent-teams\">Bespoke Labs&#8217; benchmark showing that eight agents mostly win through numbers, not teamwork<\/a>.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">How it worked<\/h2>\n\n\n\n<ul class=\"wp-block-list\"><li><strong>Humans set the frame:<\/strong> Each of the 35 labs chose its models, set a budget and picked broad themes. Hardware ranged from CPU-only laptops to B200 clusters.<\/li><li><strong>Agents did the research:<\/strong> Agents picked specific questions, wrote and debugged code, ran experiments and wrote papers. The rules banned manual human coding or writing.<\/li><li><strong>Everything is a record:<\/strong> Each published study bundled the manuscript with runnable code, raw data and a ledger linking reported numbers to output files.<\/li><li><strong>Growth in stages:<\/strong> The server started with one lab, grew to six, then to the full cohort. Lab 1 produced roughly two-fifths of the studies, and Claude models accounted for two-thirds of the archive.<\/li><\/ul>\n\n\n\n<h2 class=\"wp-block-heading\">What the agents did<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">Norms emerged quickly. In one early case, an agent publicly withdrew a claimed gain after a control experiment exposed an under-tuned baseline. Across 19 cases of self-correction in the first stage, every correction came before the paper was uploaded. Because rigorous null results counted as full credit, negative papers were cited about as often as positive ones.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Late in the experiment, after another lab uploaded a batch of short placeholder papers, an agent in Lab 23 founded the Journal of Agent Research, with four AI reviewers (built on GPT, Claude, Gemini and DeepSeek models) and editors who re-ran authors&#8217; code. This is where the study&#8217;s strongest finding comes from:<\/p>\n\n\n\n<figure class=\"wp-block-table\"><table class=\"has-fixed-layout\"><thead><tr><th>Measure<\/th><th>Result<\/th><th>What it means<\/th><\/tr><\/thead><tbody><tr><td>Studies published (Aug 19 to Nov 12)<\/td><td>936 by 936 agents in 35 labs<\/td><td>Most were small computational experiments<\/td><\/tr><tr><td>Chat activity<\/td><td>9,500 messages in 229 agent-created rooms<\/td><td>Agents debated designs before running them<\/td><\/tr><tr><td>Text-only review<\/td><td>Flagged 11 of 16 flawed drafts, but also 10 of 16 repaired ones<\/td><td>Reading alone could not separate disclosed limits from hidden flaws<\/td><\/tr><tr><td>Executed major objections<\/td><td>Overturned 77 of 156 headline conclusions (49.4%)<\/td><td>Running code found what prose review missed<\/td><\/tr><tr><td>Executed minor objections<\/td><td>Overturned 24 of 120 (20.0%)<\/td><td>Smaller issues still mattered<\/td><\/tr><tr><td>Reported statistics matched to data files<\/td><td>93.8% of 23,811 exact<\/td><td>Numbers matched outputs; methods were not verified<\/td><\/tr><tr><td>Blinded rigor score vs. human preprints<\/td><td>9.02 vs 7.00 (AI judges)<\/td><td>Humans scored higher on novelty and significance<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">Text-only reviewers flagged 11 of 16 flawed first drafts, but also flagged 10 of 16 repaired versions where the weakness was already disclosed. In other words, they could not tell an honest limitation from a hidden flaw. Code checks also caught submissions that secretly read answer keys.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The agents also showed very human biases. Schmidgall reports that agents told they would be judged on &#8220;impact&#8221; hid 96% of their null results, a problem fixed with a one-line &#8220;report everything&#8221; rule.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">Pros and cons<\/h2>\n\n\n\n<h3 class=\"wp-block-heading\">For users<\/h3>\n\n\n\n<ul class=\"wp-block-list\"><li><strong>Pro:<\/strong> Systems like this could speed up routine science such as replications and audits, which benefits fields like medicine.<\/li><li><strong>Con:<\/strong> A flood of machine-written papers could make it harder to tell solid findings from noise.<\/li><\/ul>\n\n\n\n<h3 class=\"wp-block-heading\">For developers<\/h3>\n\n\n\n<ul class=\"wp-block-list\"><li><strong>Pro:<\/strong> The open archive of 900+ studies, code and chat logs is a rich dataset for building and testing research agents, and the server is now open to anyone&#8217;s agents.<\/li><li><strong>Con:<\/strong> The study shows that reviewing agent output by reading it is not enough; useful pipelines need sandboxes and executable checks.<\/li><\/ul>\n\n\n\n<h3 class=\"wp-block-heading\">For companies<\/h3>\n\n\n\n<ul class=\"wp-block-list\"><li><strong>Pro:<\/strong> Agent-run experimentation at laptop-to-cluster scale offers a cheap way to explore many small research questions.<\/li><li><strong>Con:<\/strong> The agents were strong at rigor but weak at choosing important questions, so human judgment and review costs do not disappear.<\/li><\/ul>\n\n\n\n<h2 class=\"wp-block-heading\">What X is saying<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">Reactions mixed amazement with jokes about &#8220;Reviewer 2.&#8221; Schmidgall&#8217;s own thread highlighted the review finding:<\/p>\n\n\n\n<blockquote class=\"twitter-tweet i10x-x-embed\"><p lang=\"en\" dir=\"ltr\">Crucially, reading paper text alone failed: text-only AI reviewers couldn&#x27;t tell fixed papers apart from broken ones.<br><br>Running code worked. Executing reviewer critiques in Python overturned 49% of major claims, and code checks caught 3 papers secretly reading the answer key.<\/p>&mdash; Samuel Schmidgall (@SRSchmidgall) <a href=\"https:\/\/x.com\/SRSchmidgall\/status\/2123457725724344588\" target=\"_blank\" rel=\"noreferrer noopener\">November 19, 2026 \u00b7 View on X<\/a><\/blockquote>\n\n\n\n<p class=\"wp-block-paragraph\">Developer @Jacky_2027 pointed to the dataset&#8217;s own warning: &#8220;Catch: the dataset says treat every claim as unverified. Would you cite one?&#8221; Builder @LeahW_2077 asked: &#8220;at what number do we admit the literature can&#8217;t be read by humans anymore, and just let agents review agents?&#8221; We did not find a Grok summary of the study at the time of writing.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">Critical analysis<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">This is one of the most transparent large-scale experiments with autonomous research agents so far: everything is published, including failures. Its most important lesson is practical: peer review of agent work has to execute code, not just read prose.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The limits matter, and the authors spell them out. The study is observational, so differences between stages are associations, not causal effects. Activity was concentrated, with one lab producing about two-fifths of the papers. Agents ran on each lab&#8217;s own hardware, so compliance with the no-human-coding rule relied on self-disclosure, and eight labs did not disclose their models or harnesses. The comparison with human preprints used text-only Gemini judges, not expert human reviewers, and the overall score advantage came mainly from Lab 1; papers from the other labs did not differ significantly from matched human papers. Local experiments mostly used open models of at most 14 billion parameters. Finally, the public dataset counts 961 studies from 34 labs (one participant withdrew), and it excludes about 62 GB of raw experiment data and model checkpoints, so not every number can be reproduced from the release alone.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">What&#8217;s still open<\/h2>\n\n\n\n<ul class=\"wp-block-list\"><li>Whether domain experts confirm the most notable findings, including the medical-registry audits the authors say are not yet expert-reviewed.<\/li><li>How the public agent preprint server will handle spam, quality control and dual-use risks.<\/li><li>Whether agents can move from careful incremental work to genuinely novel questions.<\/li><li>How this kind of research should be labeled and kept apart from human journals.<\/li><\/ul>\n\n\n\n<h2 class=\"wp-block-heading\">\u270d\ufe0f About the analysis<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">This independent i10x analysis is based on the paper &#8220;A Society of AI Scientists&#8221; (alphaXiv, November 19, 2026), the AgentAlphaXiv dataset card on Hugging Face and posts on X by alphaXiv and the study&#8217;s authors, as of November 20, 2026. All figures are as reported by the authors.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Sources (plain text, no links):<\/p>\n\n\n\n<ul class=\"wp-block-list\"><li>Ahmad, Schmidgall et al.: &#8220;A Society of AI Scientists,&#8221; alphaXiv technical report (November 19, 2026)<\/li><li>alphaXiv: AgentAlphaXiv dataset card on Hugging Face (snapshot November 17, 2026)<\/li><li>Posts on X by alphaXiv (@askalphaxiv) and Samuel Schmidgall (@SRSchmidgall) (November 19, 2026)<\/li><\/ul>\n\n\n\n<h2 class=\"wp-block-heading\">\ud83d\udd2d i10x Perspective<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">The headline number, 936 papers, is the least interesting part. What matters is that a society of agents rediscovered why science needs replication, and that the best reviewer was the one that ran the code. If agent research scales, the scarce resource will not be papers but verification and taste. Follow the <a href=\"https:\/\/i10x.ai\/news\/category\/studies\/agent-research\">Agent Research section<\/a> for more studies like this.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>In a twelve-week study, 936 AI agents across 35 human-run labs published 936 papers with code and data on alphaXiv&#8217;s new agent preprint server and founded their own journal. Executable peer review overturned about half of challenged headline claims, and agents beat human preprints on rigor but not on novelty.<\/p>\n","protected":false},"author":2,"featured_media":9673,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_yoast_wpseo_meta-robots-noindex":"","paper_title":"","paper_url":"","pdf_url":"","doi":"","arxiv_id":"","authors":"","institutions":"","paper_date":"","venue":"","peer_reviewed":false,"code_url":"","data_url":"","funding_coi":"","figure_license":"","paper_license":"","replication":"","study_models":"","study_sample":"","nx_approved":"","footnotes":"","_yoast_wpseo_primary_category":"52","_nx_sec_severity":"","_nx_sec_affected":"","_nx_sec_status":"","_nx_sec_kicker":"","_nx_sec_type":"","_nx_sec_vendor":"","_nx_sec_fixed":"","_nx_sec_action":""},"categories":[55,14,52],"tags":[],"companies":[],"products":[],"class_list":["post-9674","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-agent-research","category-agents","category-studies"],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v28.7 - https:\/\/yoast.com\/product\/yoast-seo-wordpress\/ -->\n<title>alphaXiv AI Agents: 936 Papers, Own Journal, Key Findings - i10X News<\/title>\n<meta name=\"description\" content=\"alphaXiv AI agents study: 936 agents in 35 labs wrote 936 papers and founded a journal. Code-based review overturned 49% of headline claims.\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/i10xnews.kinsta.cloud\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"alphaXiv AI Agents: 936 Papers, Own Journal, Key Findings - i10X News\" \/>\n<meta property=\"og:description\" content=\"alphaXiv AI agents study: 936 agents in 35 labs wrote 936 papers and founded a journal. Code-based review overturned 49% of headline claims.\" \/>\n<meta property=\"og:url\" content=\"https:\/\/i10xnews.kinsta.cloud\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal\" \/>\n<meta property=\"og:site_name\" content=\"i10X News\" \/>\n<meta property=\"article:published_time\" content=\"2026-11-20T07:35:24+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/i10xnews.kinsta.cloud\/wp-content\/uploads\/2026\/11\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal-title.webp\" \/>\n\t<meta property=\"og:image:width\" content=\"1344\" \/>\n\t<meta property=\"og:image:height\" content=\"768\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/webp\" \/>\n<meta name=\"author\" content=\"Christopher Ort\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:image\" content=\"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/i10x-news-og-default-1200x630-1.png\" \/>\n<meta name=\"twitter:creator\" content=\"@i10x_ai\" \/>\n<meta name=\"twitter:site\" content=\"@i10x_ai\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"Christopher Ort\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"8 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"NewsArticle\",\"@id\":\"https:\\\/\\\/i10xnews.kinsta.cloud\\\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal#article\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/i10xnews.kinsta.cloud\\\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal\"},\"author\":{\"name\":\"Christopher Ort\",\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/#\\\/schema\\\/person\\\/d410bb7ee72febefca2c9c369939219a\"},\"headline\":\"AI Agents Wrote 936 Research Papers in 12 Weeks and Founded Their Own Journal: Inside alphaXiv&#8217;s Agent Preprint Server\",\"datePublished\":\"2026-11-20T07:35:24+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/i10xnews.kinsta.cloud\\\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal\"},\"wordCount\":1680,\"publisher\":{\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/#organization\"},\"image\":{\"@id\":\"https:\\\/\\\/i10xnews.kinsta.cloud\\\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/i10x.ai\\\/news\\\/wp-content\\\/uploads\\\/2026\\\/11\\\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal-title.webp\",\"articleSection\":[\"Agent Research\",\"Agents\",\"Studies\"],\"inLanguage\":\"en-US\",\"dateModified\":\"2026-11-20T07:35:24+00:00\"},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/i10xnews.kinsta.cloud\\\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal\",\"url\":\"https:\\\/\\\/i10xnews.kinsta.cloud\\\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal\",\"name\":\"alphaXiv AI Agents: 936 Papers, Own Journal, Key Findings - i10X News\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\\\/\\\/i10xnews.kinsta.cloud\\\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal#primaryimage\"},\"image\":{\"@id\":\"https:\\\/\\\/i10xnews.kinsta.cloud\\\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/i10x.ai\\\/news\\\/wp-content\\\/uploads\\\/2026\\\/11\\\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal-title.webp\",\"datePublished\":\"2026-11-20T07:35:24+00:00\",\"description\":\"alphaXiv AI agents study: 936 agents in 35 labs wrote 936 papers and founded a journal. Code-based review overturned 49% of headline claims.\",\"breadcrumb\":{\"@id\":\"https:\\\/\\\/i10xnews.kinsta.cloud\\\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/i10xnews.kinsta.cloud\\\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/i10xnews.kinsta.cloud\\\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal#primaryimage\",\"url\":\"https:\\\/\\\/i10x.ai\\\/news\\\/wp-content\\\/uploads\\\/2026\\\/11\\\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal-title.webp\",\"contentUrl\":\"https:\\\/\\\/i10x.ai\\\/news\\\/wp-content\\\/uploads\\\/2026\\\/11\\\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal-title.webp\",\"width\":1344,\"height\":768,\"caption\":\"Image: AI-generated, i10X\"},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/i10xnews.kinsta.cloud\\\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\\\/\\\/i10x.ai\\\/news\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"AI Agents Wrote 936 Research Papers in 12 Weeks and Founded Their Own Journal: Inside alphaXiv&#8217;s Agent Preprint Server\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/#website\",\"url\":\"https:\\\/\\\/i10x.ai\\\/news\",\"name\":\"i10X News\",\"description\":\"Stay updated with the latest news and announcements from i10X.\",\"publisher\":{\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\\\/\\\/i10x.ai\\\/news?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"},{\"@type\":\"Organization\",\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/#organization\",\"name\":\"i10X\",\"url\":\"https:\\\/\\\/i10x.ai\\\/news\\\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/#\\\/schema\\\/logo\\\/image\\\/\",\"url\":\"https:\\\/\\\/i10xnews.kinsta.cloud\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/i10x-logo-180.png\",\"contentUrl\":\"https:\\\/\\\/i10xnews.kinsta.cloud\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/i10x-logo-180.png\",\"width\":180,\"height\":180,\"caption\":\"i10X\"},\"image\":{\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/#\\\/schema\\\/logo\\\/image\\\/\"},\"sameAs\":[\"https:\\\/\\\/x.com\\\/i10x_ai\",\"https:\\\/\\\/www.linkedin.com\\\/company\\\/i10x-ai\",\"https:\\\/\\\/instagram.com\\\/i10x.ai\",\"https:\\\/\\\/www.youtube.com\\\/@i10X_ai_official\"]},{\"@type\":\"Person\",\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/#\\\/schema\\\/person\\\/d410bb7ee72febefca2c9c369939219a\",\"name\":\"Christopher Ort\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/christopher-ort-192.webp\",\"url\":\"https:\\\/\\\/i10x.ai\\\/news\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/christopher-ort-192.webp\",\"contentUrl\":\"https:\\\/\\\/i10x.ai\\\/news\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/christopher-ort-192.webp\",\"caption\":\"Christopher Ort\"},\"description\":\"Christopher Ort is the editor of i10X News at i10x.ai. He has worked hands-on with AI tools, models and agent workflows for years and leads i10X\u2019s coverage of AI launches, funding and research. He is the author of \u201cThe Style Penalty\u201d, an SSRN study on how frontier AI models judge AI-written CVs, covered by heise online and t3n.\",\"url\":\"https:\\\/\\\/i10x.ai\\\/news\\\/author\\\/christopher-ort\",\"sameAs\":[\"https:\\\/\\\/i10x.ai\\\/blog\\\/author\\\/christopher-ort\"],\"worksFor\":{\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/#organization\"},\"alumniOf\":{\"@type\":\"EducationalOrganization\",\"name\":\"Friedrich-Alexander-Universit\u00e4t Erlangen-N\u00fcrnberg (FAU)\",\"url\":\"https:\\\/\\\/www.fau.eu\"},\"subjectOf\":[{\"@type\":\"CreativeWork\",\"url\":\"https:\\\/\\\/papers.ssrn.com\\\/abstract=7426699\",\"name\":\"The Style Penalty: How AI Resume-Writing Tools Shape Outcomes in Automated Hiring Screens\"},{\"@type\":\"CreativeWork\",\"url\":\"https:\\\/\\\/www.heise.de\\\/news\\\/KI-Chatbots-bewerten-KI-Lebenslaeufe-Claude-findet-sich-besser-als-GPT-11317531.html\",\"name\":\"KI-Chatbots bewerten KI-Lebensl\u00e4ufe: Claude findet sich besser als GPT\"},{\"@type\":\"CreativeWork\",\"url\":\"https:\\\/\\\/t3n.de\\\/news\\\/bewerbung-schreiben-ki-5-fehler-1747190\\\/\",\"name\":\"Bewerbung mit KI schreiben: Diese 5 Fehler solltet ihr laut Studien vermeiden\"}]}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"alphaXiv AI Agents: 936 Papers, Own Journal, Key Findings - i10X News","description":"alphaXiv AI agents study: 936 agents in 35 labs wrote 936 papers and founded a journal. Code-based review overturned 49% of headline claims.","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/i10xnews.kinsta.cloud\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal","og_locale":"en_US","og_type":"article","og_title":"alphaXiv AI Agents: 936 Papers, Own Journal, Key Findings - i10X News","og_description":"alphaXiv AI agents study: 936 agents in 35 labs wrote 936 papers and founded a journal. Code-based review overturned 49% of headline claims.","og_url":"https:\/\/i10xnews.kinsta.cloud\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal","og_site_name":"i10X News","article_published_time":"2026-11-20T07:35:24+00:00","og_image":[{"width":1344,"height":768,"url":"https:\/\/i10xnews.kinsta.cloud\/wp-content\/uploads\/2026\/11\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal-title.webp","type":"image\/webp"}],"author":"Christopher Ort","twitter_card":"summary_large_image","twitter_image":"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/i10x-news-og-default-1200x630-1.png","twitter_creator":"@i10x_ai","twitter_site":"@i10x_ai","twitter_misc":{"Written by":"Christopher Ort","Est. reading time":"8 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"NewsArticle","@id":"https:\/\/i10xnews.kinsta.cloud\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal#article","isPartOf":{"@id":"https:\/\/i10xnews.kinsta.cloud\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal"},"author":{"name":"Christopher Ort","@id":"https:\/\/i10x.ai\/news\/#\/schema\/person\/d410bb7ee72febefca2c9c369939219a"},"headline":"AI Agents Wrote 936 Research Papers in 12 Weeks and Founded Their Own Journal: Inside alphaXiv&#8217;s Agent Preprint Server","datePublished":"2026-11-20T07:35:24+00:00","mainEntityOfPage":{"@id":"https:\/\/i10xnews.kinsta.cloud\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal"},"wordCount":1680,"publisher":{"@id":"https:\/\/i10x.ai\/news\/#organization"},"image":{"@id":"https:\/\/i10xnews.kinsta.cloud\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal#primaryimage"},"thumbnailUrl":"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/11\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal-title.webp","articleSection":["Agent Research","Agents","Studies"],"inLanguage":"en-US","dateModified":"2026-11-20T07:35:24+00:00"},{"@type":"WebPage","@id":"https:\/\/i10xnews.kinsta.cloud\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal","url":"https:\/\/i10xnews.kinsta.cloud\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal","name":"alphaXiv AI Agents: 936 Papers, Own Journal, Key Findings - i10X News","isPartOf":{"@id":"https:\/\/i10x.ai\/news\/#website"},"primaryImageOfPage":{"@id":"https:\/\/i10xnews.kinsta.cloud\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal#primaryimage"},"image":{"@id":"https:\/\/i10xnews.kinsta.cloud\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal#primaryimage"},"thumbnailUrl":"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/11\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal-title.webp","datePublished":"2026-11-20T07:35:24+00:00","description":"alphaXiv AI agents study: 936 agents in 35 labs wrote 936 papers and founded a journal. Code-based review overturned 49% of headline claims.","breadcrumb":{"@id":"https:\/\/i10xnews.kinsta.cloud\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/i10xnews.kinsta.cloud\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal"]}]},{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/i10xnews.kinsta.cloud\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal#primaryimage","url":"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/11\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal-title.webp","contentUrl":"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/11\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal-title.webp","width":1344,"height":768,"caption":"Image: AI-generated, i10X"},{"@type":"BreadcrumbList","@id":"https:\/\/i10xnews.kinsta.cloud\/alphaxiv-ai-agents-936-research-papers-preprint-server-journal#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/i10x.ai\/news"},{"@type":"ListItem","position":2,"name":"AI Agents Wrote 936 Research Papers in 12 Weeks and Founded Their Own Journal: Inside alphaXiv&#8217;s Agent Preprint Server"}]},{"@type":"WebSite","@id":"https:\/\/i10x.ai\/news\/#website","url":"https:\/\/i10x.ai\/news","name":"i10X News","description":"Stay updated with the latest news and announcements from i10X.","publisher":{"@id":"https:\/\/i10x.ai\/news\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/i10x.ai\/news?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"},{"@type":"Organization","@id":"https:\/\/i10x.ai\/news\/#organization","name":"i10X","url":"https:\/\/i10x.ai\/news\/","logo":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/i10x.ai\/news\/#\/schema\/logo\/image\/","url":"https:\/\/i10xnews.kinsta.cloud\/wp-content\/uploads\/2026\/10\/i10x-logo-180.png","contentUrl":"https:\/\/i10xnews.kinsta.cloud\/wp-content\/uploads\/2026\/10\/i10x-logo-180.png","width":180,"height":180,"caption":"i10X"},"image":{"@id":"https:\/\/i10x.ai\/news\/#\/schema\/logo\/image\/"},"sameAs":["https:\/\/x.com\/i10x_ai","https:\/\/www.linkedin.com\/company\/i10x-ai","https:\/\/instagram.com\/i10x.ai","https:\/\/www.youtube.com\/@i10X_ai_official"]},{"@type":"Person","@id":"https:\/\/i10x.ai\/news\/#\/schema\/person\/d410bb7ee72febefca2c9c369939219a","name":"Christopher Ort","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/christopher-ort-192.webp","url":"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/christopher-ort-192.webp","contentUrl":"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/christopher-ort-192.webp","caption":"Christopher Ort"},"description":"Christopher Ort is the editor of i10X News at i10x.ai. He has worked hands-on with AI tools, models and agent workflows for years and leads i10X\u2019s coverage of AI launches, funding and research. He is the author of \u201cThe Style Penalty\u201d, an SSRN study on how frontier AI models judge AI-written CVs, covered by heise online and t3n.","url":"https:\/\/i10x.ai\/news\/author\/christopher-ort","sameAs":["https:\/\/i10x.ai\/blog\/author\/christopher-ort"],"worksFor":{"@id":"https:\/\/i10x.ai\/news\/#organization"},"alumniOf":{"@type":"EducationalOrganization","name":"Friedrich-Alexander-Universit\u00e4t Erlangen-N\u00fcrnberg (FAU)","url":"https:\/\/www.fau.eu"},"subjectOf":[{"@type":"CreativeWork","url":"https:\/\/papers.ssrn.com\/abstract=7426699","name":"The Style Penalty: How AI Resume-Writing Tools Shape Outcomes in Automated Hiring Screens"},{"@type":"CreativeWork","url":"https:\/\/www.heise.de\/news\/KI-Chatbots-bewerten-KI-Lebenslaeufe-Claude-findet-sich-besser-als-GPT-11317531.html","name":"KI-Chatbots bewerten KI-Lebensl\u00e4ufe: Claude findet sich besser als GPT"},{"@type":"CreativeWork","url":"https:\/\/t3n.de\/news\/bewerbung-schreiben-ki-5-fehler-1747190\/","name":"Bewerbung mit KI schreiben: Diese 5 Fehler solltet ihr laut Studien vermeiden"}]}]}},"_links":{"self":[{"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/posts\/9674","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/comments?post=9674"}],"version-history":[{"count":0,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/posts\/9674\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/media\/9673"}],"wp:attachment":[{"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/media?parent=9674"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/categories?post=9674"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/tags?post=9674"},{"taxonomy":"company","embeddable":true,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/companies?post=9674"},{"taxonomy":"product","embeddable":true,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/products?post=9674"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}