OpenAI Shares AI Discoveries in Math, CS, and Physics
OpenAI has announced three new discoveries produced by its internal models in mathematics, computer science, and physics, while experimenting with collaborative academic release strategies.
Radwan AbdulhadiEditor and technology writer
New Scientific Discoveries From Internal Models
OpenAI has announced three new discoveries produced by its internal models spanning mathematics, computer science, and physics. The findings focus on harmonic analysis, complexity theory, and quantum field theory. According to the organization, these manuscripts are treated as raw research data, requiring new approaches to help academic communities convert them into scientific understanding.
To handle these releases, OpenAI reached out to external experts to act as shepherds who organized groups to understand the new results. Different release strategies were applied depending on the scientific field, reflecting feedback from the associated academic communities.
Collaborative Principles and Release Strategies
The external collaborations follow four core principles: transparency, no academic credit for OpenAI's internal models, recognition for the working groups, and complete independence for participating researchers. Participants are not required to sign nondisclosure agreements, allowing them to discuss the work freely and publish their conclusions.
For the mathematics result, OpenAI released the AI output directly to let researchers organize their own efforts. For the computer science result, an external group of collaborators was formed while making the AI output publicly available. For the physics result, the organization deferred entirely to an external group of collaborators regarding the release of materials.
Mathematics Breakthrough in Harmonic Analysis
The mathematics discovery concerns harmonic analysis and the local smoothing conjecture proposed by Sogge in 1991. Previous milestones include proof in two spatial dimensions by Guth, Wang, and Zhang in 2019, and the proof of the Kakeya conjecture in three dimensions by Wang and Zahl in 2025.
OpenAI's models produced a proof of the local smoothing conjecture in every dimension, thereby resolving the Kakeya conjecture and Stein’s restriction conjecture in all dimensions using a novel auxiliary partial differential equation. Following expert recommendations, OpenAI released both an AI natural language output writeup and a Lean proof in its public GitHub repository, adopting a strategy similar to its previous October 6th release.
Computer Science and Complexity Theory Findings
In computer science, a separate collaboration focused on a result concerning the Exponential Time Hypothesis (ETH). The ETH is a longstanding conjecture about the computational complexity of Boolean formula satisfiability, a problem closely related to the P vs. NP problem. Widely believed to be true, the hypothesis has underpinned hundreds of papers across graph theory, computational biology, databases, and computational geometry.
OpenAI's system recently proved that the ETH is false, a result subsequently verified in Lean. This refutation reopens fundamental questions regarding how much faster computational tasks can become and may guide the search for superior algorithms moving forward.
Sources
- OpenAI NewsWorking with academic communities to build scientific understanding
Continue chronologically
Related entity coverage
Related stories
OpenAI Shares AI Progress in Mathematics and Lean Proofs
OpenAI publishes new mathematical results from an internal frontier model, including Lean proof formalizations and research details on GitHub.
OpenAI Announces New Math, Physics and Computer Science Discoveries
OpenAI unveils new discoveries in math, physics, and computer science, including models generating proof that the Exponential Time Hypothesis is false.
OpenAI Releases Batch of Mathematical Breakthroughs on GitHub
OpenAI has published 722 manuscripts detailing solutions to open math problems, drawing mixed reactions from the mathematical community.
OpenAI Announces Three Major Science Discoveries with Academics
OpenAI unveils three new discoveries in math, computer science, and physics, working alongside external academic experts and advisors.
