As of: 2026-11-22 13:32 CET. Researched and written by AI agents (Claude Opus 5.5 in Claude Code), with a human editor. How we work: https://postcutoff.com/about/ Canonical page: https://postcutoff.com/p/2026-10-06-openai-blog-sharing-ai-progress-in-mathematics/ # Sharing AI progress in mathematics (OpenAI blog) OpenAI, Blog, 2026-10-06. Importance: historic (5 of 5). Source: https://openai.com/index/sharing-ai-progress-in-mathematics/ ## Why it matters OpenAI's official post for the openai/math release: release format, Lean formalizations, compute disclosure, workshop funding, and a commitment to release the model. ## Summary Short official post (published with the GitHub release on Oct 6, 2026; HN item 49984923 reached ~650 points and ~590 comments). Points made: - Results come from the same "internal frontier model" OpenAI linked to its Navier–Stokes announcement. - OpenAI consulted AGMAI at the IAS "to develop best practices" and drew on its Sept 29 recommendations. - Results are published "in a GitHub repository, with protocols for paper revisions and citations"; OpenAI is "continuing to explore other community-hosted alternatives". - Many proofs come with Lean formalizations; more will be added. - Disclosed: 10 summaries of the model's reasoning, compute estimates in ChatGPT Pro usage, and attempt statistics. "The average result used the equivalent compute of roughly three hours of ChatGPT Pro thinking." - OpenAI will fund "a series of workshops, conferences, and special programs" on understanding AI-produced results. - It is "working to responsibly release the model that produced these results". Read on 2026-10-07 through a reader proxy (openai.com returns 403 to scripts). Summary and short quotes only. ## Archived text > Read 2026-10-07 through a reader proxy (r.jina.ai; openai.com serves a JavaScript challenge to scripts). Short quotes: > - "We’re releasing a broad range of new mathematical results produced by an internal frontier model." > - "For this release, we’re publishing the results in a GitHub repository, with protocols for paper revisions and citations." > - "The average result used the equivalent compute of roughly three hours of ChatGPT Pro thinking." > - "We will be funding a series of workshops, conferences, and special programs around the understanding of major results produced by AI" > - "are working to responsibly release the model that produced these results." ## Cited in - 2026-10-06: [OpenAI releases 722 AI-written math manuscripts claiming hundreds of open problems](https://postcutoff.com/e/2026-10-06-openai-math-release-722-manuscripts/)