AI · Oct 21, 2026
OpenAI on Google Cloud lets companies spend Google commitments on GPT-6, and that changes procurementMathematicians object as OpenAI reportedly plans a GitHub release of hundreds of open-problem results
· Updated · Signal Desk newsroom
Analysis of reporting by WIRED. How we work

WIRED reports OpenAI told people it plans to post hundreds of results on long-standing math problems to GitHub, after about 40 mathematicians asked in August for proper papers. OpenAI says a new model has resolved more than 100 open problems and that no release time is set.
Key takeaways
- OpenAI reportedly plans to post hundreds of math results to GitHub, per WIRED.
- About 40 mathematicians met with OpenAI in August and asked for papers instead.
- OpenAI says a model that began training August 28 resolved more than 100 open problems plus Navier–Stokes.
- Spokesperson Lindsay McCallum says no release time has been set and an IAS advisory group is involved.
- Accusations of front-running and aggressive conduct are disputed by OpenAI and Sébastien Bubeck.
What WIRED reported
According to WIRED, OpenAI has told people it plans to post hundreds of results on long-standing unsolved math problems to GitHub. That follows an August meeting where roughly 40 mathematicians asked the company to publish proper papers instead of blog posts or social media announcements (the consultation). Northwestern mathematician Bryna Kra, who attended, says that input was apparently ignored.
OpenAI spokesperson Lindsay McCallum says a new internal model, which began training August 28, has resolved more than 100 long-standing open problems across most areas of mathematics, in addition to the Navier–Stokes Millennium Prize problem. She says the company is working to release results responsibly with input from an advisory group at the Institute for Advanced Study, formed in mid-September, and that no release time has been set. None of the claimed resolutions had been independently verified in WIRED’s reporting.
Why format matters
If the results hold up, they would rank among the most significant contributions AI has made to human knowledge. The frustration is not about whether AI can do math; Kra says she is glad to have a powerful new tool. Her request is that results arrive in a form mathematicians can verify, attribute and build on.
Papers explain methods, cite prior work and let others check the logic. A bulk repository, however impressive, shifts the cost of checking and digesting results onto a community that did not agree to carry it. Kra told WIRED that math by press release is not how to nurture the ecosystem these models were trained on. Earlier episodes, including ten problems announced by blog and social posts in August, drew similar complaints (earlier criticism, attribution questions).
Disputes over conduct
WIRED’s reporting includes sharper allegations. NYU’s Tristan Buckmaster accuses OpenAI of front-running work he was doing with Anthropic’s Levent Alpöge after OpenAI deployed thousands of agents on a Millennium Prize problem in September. Meeting notes seen by WIRED describe OpenAI researcher Sébastien Bubeck saying he did not have to be nice; Bubeck has publicly denied asking that Alpöge be dropped as an author. NYU mathematician Nestor Guillen says colleagues perceive “mobster behavior” from the labs, a characterization OpenAI disputes. These accounts concern private conversations that cannot be independently settled from the reporting.
The issue is not OpenAI’s alone. WIRED notes mathematicians feel the field has become a showcase for both OpenAI and Anthropic as they approach IPOs, and Alpöge himself announced a result by tweet. Established routes exist: Hexagon and Palomar are designed for AI-generated and machine-verified math, and more than 4,000 mathematicians have signed the Leiden declaration setting out standards. Using them would cost labs little and build trust. More in our OpenAI coverage, Anthropic coverage and research coverage.
Who feels it
- Mathematicians
- A large unexplained batch would divert time to verification and attribution.
- Early-career researchers
- Uncertainty about credit and careers intensifies.
- AI labs
- Disclosure of scientific results becomes a reputational and governance issue ahead of IPOs.
What to watch
- Whether results arrive as papers, via Hexagon or Palomar, or as a bare repository.
- Public guidance from the IAS advisory group and whether OpenAI follows it.
- Independent verification of the Navier–Stokes and other claims.
- Whether Anthropic adopts different disclosure norms.
FAQ
What does OpenAI reportedly plan?
To post hundreds of math results to GitHub, according to WIRED.
What did mathematicians ask for?
Proper papers rather than blog posts or social media announcements.
What does OpenAI claim?
A model that began training August 28 resolved more than 100 open problems plus Navier–Stokes, per spokesperson Lindsay McCallum.
Has a release date been set?
No, OpenAI says no release time has been set.
Are the conduct allegations confirmed?
No. Bubeck has denied asking that Alpöge be dropped as an author, and OpenAI disputes the characterizations.
Companies: OpenAI
More on OpenAI
AI · Oct 14, 2026
White Circle says newer OpenAI and Google image models are easier to jailbreak than their predecessorsAI · Oct 19, 2026
Microsoft ships Guardian Agents, saying frontier models ‘can’t be trusted’AI · Oct 13, 2026
After months off the frontier, Google ships Gemini 4 Argon with benchmark claims and aggressive pricingAI · Sep 22, 2026
Opus 5.5 hands risky cyber prompts to an older Claude model, and that routing is the real launch story