- OpenAI's model generates mathematical proofs that challenge theories in harmonic analysis and computer science.
- The model refutes the Exponential Time Hypothesis, potentially impacting assumptions across multiple research fields.
- The existence proof for quantum field theory involves four-dimensional gauge fields and fermionic interactions.
- OpenAI will fund a conference to assess the implications and significance of its research.
OpenAI states that its model has generated mathematical proofs that challenge established theories in harmonic analysis and computer science. This work also includes existence proofs for quantum field theory, covering quantum chromodynamics (QCD). Researchers will further examine these findings to verify their impact and explain their significance to the broader scientific community. The company has adopted different release strategies for three projects, making mathematical and computational proofs public, while an external physics team will prepare an independent report on its research.
The mathematical results challenge established theories. One of the proofs addresses the Local Smoothness Conjecture, a problem in harmonic analysis proposed by Christopher Sogge in 1991. This conjecture describes the behavior of waves when concentrated, suggesting that extreme focusing cannot be sustained for long periods. OpenAI claims its model has proven this conjecture in every spatial dimension. According to the company, this result also resolves the Kakeya Conjecture and Stein's restriction conjecture. Previous research has established significant milestones in this field.
In 2019, Larry Guth, Hong Wang, and Ruixiang Zhang proved the Local Smoothness Conjecture in two spatial dimensions. Wang and Zahl subsequently proved the three-dimensional Kakeya Conjecture in 2025. OpenAI generated new proofs on October 18, 2026, and formalized them in Lean the following day. Lean is a proof assistant that helps mathematicians verify arguments through computer checks.
The company has released natural language proofs and their formal counterparts on GitHub.
New Challenges in Computational Theory
Computational theory is facing a new round of scrutiny. OpenAI also states that its model refutes the Exponential Time Hypothesis (ETH), a longstanding conjecture in computational complexity. This hypothesis concerns the difficulty of solving the Boolean satisfiability problem and underpins conditional hardness results in multiple research fields. Researchers have used ETH in hundreds of papers covering graph theory, computational biology, databases, and computational geometry. If the new results withstand scrutiny, they may compel researchers to reassess the assumptions underlying these studies. These models generated the refutation on October 16, 2026, and were formalized in Lean the following day.
However, refuting ETH does not automatically provide faster algorithms for existing computational problems. Researchers still need to determine the implications of this result for practical computation. Russell Impagliazzo and Ramamohan Paturi from the University of California, San Diego, will collaborate with Ryan Williams from MIT to review the findings. OpenAI has agreed to fund a conference to examine the statement and release its proof materials for independent evaluation.
Mathematical Foundations of Quantum Field Theory
The proofs for quantum field theory are still under review. This physical work involves the mathematical foundations of quantum field theory. OpenAI states that its model has generated existence proofs involving four-dimensional gauge fields and fermionic interaction theories. The results include QCD, which describes the strong force that governs interactions between quarks and gluons. This work also involves pure Yang-Mills theory, a key component of the Millennium Prize Problems from the Clay Mathematics Institute. This problem concerns the mathematical existence of quantum Yang-Mills theory and the existence of a mass gap.
Physicist David Gross will lead an external team that includes Michael Douglas, Arthur Jaffe, Igor Klebanov, and Nikita Nekrasov. The team plans to develop its own interpretations before releasing the physics materials generated by the model. OpenAI states that these three collaborations will operate independently, and participants are free to publish their findings. The company has also committed to supporting the Simons Foundation with previously announced funding of $5 million.
This funding will help researchers organize workshops and conferences to review the work and assess its broader implications.
OpenAI's Impact on Mathematics and Physics
OpenAI's research not only challenges traditional mathematical theories but may also disrupt fundamental assumptions in computational theory. The refutation of the Exponential Time Hypothesis, if validated, will force researchers to reconsider widely used assumptions in computational complexity. Furthermore, the existence proofs for quantum field theory are crucial for understanding interactions between fundamental particles, potentially opening new directions for future physics research. As these findings undergo further scrutiny, the scientific community will closely monitor their potential impact.

