The comments follow months of criticism from researchers over how OpenAI announces its results. This month alone the company has released hundreds of mathematical results, often with brief writeups that mathematicians describe as poorly written and hard to follow, leaving them to do the painstaking work of verification and interpretation. Critics cite inadequate attribution, little context and insufficient verification.

There is a tension in Altman's position, though. Following announcements of three claimed breakthroughs in mathematics, physics and computer science, researchers told The Verge they are frustrated by what they see as OpenAI's gatekeeping. Details of the claimed Yang-Mills theory result, which concerns the mathematical foundations of modern particle physics, have not been released publicly. OpenAI says that is on the advice of a hand-picked expert group, which runs counter to the advice of a previous group of mathematicians it consulted.

Altman also hinted there is much more in the pipeline. Asked whether OpenAI holds more than the hundreds of proofs already released, he replied: "We do. Of course." He admitted the effort briefly "got out of control" as employees threw the hardest problems they could at the models, and estimated the company has spent in the order of hundreds of millions of dollars on mathematical problems, including around $10 million on its Navier-Stokes solution, one of the Millennium Prize problems.