OpenAI’s New Academic Collaborations
OpenAI has initiated a collaborative effort with academic communities after its internal models produced significant research results in mathematics, computer science, and physics. The company outlined distinct release strategies for each result and plans to provide materials to external collaborators for independent analysis and explanation.
Release Strategies and Principles
OpenAI has categorized the outputs of its models as raw research data and consulted external experts for guidance on the release process. The collaborations will adhere to four guiding principles: Transparency, No Academic Credit for OpenAI, Recognition for participating groups, and Independence of volunteer participants. Participants will not be required to sign nondisclosure agreements and will have the freedom to publish their findings.
Research Results and Collaborators
Three key results have emerged from OpenAI’s models, each with designated academic stewards:
Mathematics: Local Smoothing and Kakeya Conjectures
OpenAI’s models produced a proof of the local smoothing conjecture across all dimensions, which also resolves the Kakeya conjecture and Stein’s restriction conjecture. The AI output was generated on October 18, 2026, and was formalized in Lean the following day. Both the natural language writeup and the Lean proof have been published in OpenAI’s public GitHub repository.
Computer Science: Exponential Time Hypothesis
The models also generated a refutation of the Exponential Time Hypothesis (ETH). This result was verified in Lean after the AI discovered the natural language disproof on October 16, 2026, followed by formalization on October 17, 2026. A group led by Russell Impagliazzo (UC San Diego), Ramamohan Paturi (UC San Diego), and Ryan Williams (MIT) will oversee the verification and explanation of this result. The collaborators have requested that raw data be made public and that OpenAI fund a conference to explore the implications of this work, to which OpenAI has agreed.
Physics: Yang-Mills Existence and Quantum Chromodynamics
In the realm of physics, OpenAI’s models produced natural language proofs for the existence of Yang-Mills theory on September 9, 2026, and for Quantum Chromodynamics (QCD) on September 18, 2026. These results are part of a broader effort, and the group, led by David Gross along with Michael Douglas, Arthur Jaffe, Igor Klebanov, and Nikita Nekrasov, has requested that the public release of AI-generated artifacts be delayed while they prepare their own write-up of the findings. OpenAI has complied with this request. The group has also requested that all communications regarding this effort be directed to YANGMILLS4D@gmail.com.
Supporting Academic Research
OpenAI is committed to supporting the academic community in understanding and building upon these results. This initiative complements a recent $5 million contribution to the Simons Foundation, which will fund workshops, conferences, and programs aimed at helping researchers examine and expand on the published materials.
Original source: OpenAI News