OpenAI's breakthrough in mathematics: 377 new results

OpenAI shared a comprehensive new study that showcases the capabilities of artificial intelligence in mathematics with the public. The company has obtained results in various fields of mathematics and theoretical computer science.to 377 new result familiespublished hundreds of related studies.

According to the company's announcement on October 6, the results cover distinct areas such as algebra, number theory, mathematical logic, topology, and theoretical computer science. OpenAI stated that the majority of the results were produced by an internal advanced model that has not yet been released to the public.

Hours of Computation for Each Result

According to the information shared by OpenAI, each result required an operation roughly equivalent to about three hours of ChatGPT Pro computing power.

During the model evaluation process, approximatelywith 4,000 math problemswere confronted. The resulting studies were later classified by importance levels and turned into result families and academic papers.

The company stated that the mathematical proofs of a significant portion of the obtained results were checked using the formal verification language Lean to enable computer verification. Thus, not only the result produced by the AI but also the logical consistency of the proof was inspected by the computer.

The Role of Artificial Intelligence in Mathematics is Growing

OpenAI emphasizes that the new results are not only solutions to existing problems but also a sign of a new era for how AI can be used in scientific research.

The company plans to hold workshops and conferences to better understand the mathematical work produced by artificial intelligence, and also to support specialized programs in this field.

OpenAI also published summaries for ten different problems on how the model arrived at its conclusions. This aims to allow researchers to examine not only the results but also the AI's problem‑solving process.

It Came After the Navier‑Stokes Breakthrough

OpenAI's latest announcement came after the company's Navier‑Stokes study, which made a big splash in the mathematics world in September.

The company, on September 8, 2026, has been unable to solve for about 90 yearsNavier‑Stokes Millennium Prize Problemhad published a solution generated by artificial intelligence. According to OpenAI’s statement, the study prepared by an internal system shows that the equations describing fluid motion can develop a singularity in finite time under certain conditions. Along with the mathematical explanation of the solution, a proof formalized with Lean was also shared.

Results Opened for Review by the Scientific Community

OpenAI said the new studies are now available on GitHub and that corrections and updates to the papers will also be tracked publicly.

It was stated that the company’s Mathematics and Artificial Intelligence Advisory Group also played a role in shaping this publication process. OpenAI announced that it will continue to improve publication standards to share similar scientific results more transparently and understandably in the future.

This performance of AI in mathematics is reigniting discussions about computers moving beyond being mere tools that apply known methods, to actively playing a role in producing new mathematical results.