OpenAI Claims Three Theory Breakthroughs, Says ETH Is False

On October 26, 2026, OpenAI published model results on problems including the Kakeya conjecture, the Exponential Time Hypothesis, and quantum chromodynamics, along with an account of how researchers could verify and build on them. The announcement did not say which model was used, and it did not say whether the results had been independently reviewed.

On October 27, 2026, Techmeme and The Next Web reported that OpenAI had announced three new discoveries spanning mathematics, computer science, and physics, covering harmonic analysis, complexity theory, and quantum field theory. One of the claims was that the Exponential Time Hypothesis (ETH) is false. According to The Next Web, OpenAI said that an internal model had refuted ETH, a conjecture that the company described as widely believed and one on which hundreds of computer science papers rest, and that OpenAI announced two more results that same day, one in mathematics and one in physics. All of these claims were published by OpenAI itself, and there has so far been no third-party verification of any of them. The status of each result therefore rests on OpenAI's own publication and the verification process it described, rather than on independent confirmation by outside researchers.