{"version":1,"type":"story","url":"https://digestai.news/story/openai-releases-722-math-papers-from-unreleased-model","json":"https://digestai.news/story/openai-releases-722-math-papers-from-unreleased-model.json","markdown":"https://digestai.news/story/openai-releases-722-math-papers-from-unreleased-model.md","slug":"openai-releases-722-math-papers-from-unreleased-model","headline":"OpenAI releases 722 math papers from unreleased model","summary":"OpenAI published 722 math papers from an unreleased internal model, grouped into 372 result families. The boldest claim is a proof of the quasi-Riemann hypothesis, a weaker version of the $1 million Millennium Prize problem. Out of these, 162 papers have results written in Lean, a software that verifies every step of the proof process automatically.\n\nThe company claims nearly all results took just one prompt and averaged 3 hours of ChatGPT Pro compute. This contrasts with last month’s Navier–Stokes proof, which required 10,000 AI agents. Levent Alpöge, a mathematician at rival Anthropic, called it the most significant moment in mathematical history. Critics argue OpenAI released too many results too quickly, breaking earlier promises to space them out.","keyPoints":["OpenAI released 722 math papers from an unreleased internal model, grouped into 372 result families","162 papers use Lean software for automated proof verification, with results averaging 3 hours of compute per prompt","Levent Alpöge, an Anthropic mathematician, called it the most significant moment in mathematical history"],"whyItMatters":"The release shows math problems can now scale with compute rather than human effort, accelerating progress in fields where proofs are computationally intensive. It also reignites debates about transparency and speed in AI-driven research.","category":{"slug":"research","name":"Research","url":"https://digestai.news/category/research"},"entities":{"companies":["OpenAI","Anthropic","Mistral","Reflection AI","Figure AI","Google"],"models":["GPT-6 Astra","Claude Opus 5.5","Le Chonk","Beam","Kimi K3","Glean"],"people":["Levent Alpöge","Brett Adcock","Abidur Chowdhury"]},"firstPublishedAt":"2026-10-07T10:00:00Z","updatedAt":"2026-10-07T10:00:00Z","sourceCount":1,"hasPrimarySource":false,"sources":[{"outlet":"The Rundown AI","title":"OpenAI's math avalanche continues","url":"https://therundown.ai/articles/openai-math-avalanche-continues","publishedAt":"2026-10-07T10:00:00Z","type":"newsletter","primary":false,"lead":true}],"sourceNotes":null,"discussions":[],"thread":null,"cite":{"text":"Digest AI, \"OpenAI releases 722 math papers from unreleased model\", 7 October 2026, https://digestai.news/story/openai-releases-722-math-papers-from-unreleased-model","publisher":"Digest AI","title":"OpenAI releases 722 math papers from unreleased model","datePublished":"2026-10-07T10:00:00Z","url":"https://digestai.news/story/openai-releases-722-math-papers-from-unreleased-model"},"generatedBy":"Written by Digest AI's editorial model from the linked sources; the sources are the record.","license":"Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse"}