OpenAI Claims $1M Math Proof; Mathematician Alleges Intimidation

Buckmaster's core suspicion is about data. "I asked whether the model had been trained on, or had access to, our sessions in Codex, into which we had been putting all our drafts for the whole of this project," he wrote. "I was told the model did not look up user data. I asked again, about training, and I did not get an answer."

He also says the approach itself was too obscure to be a coincidence: "Almost nobody else I know of was working on it. It is not the direction one arrives at in a few days by giving a model the problem statement."

Buckmaster says Bubeck floated a compromise in which Buckmaster could publish and claim the prize — but only if he said OpenAI's model had also solved the problem, and only if he removed Alpöge's name from the paper, because OpenAI disliked his Anthropic affiliation. When Buckmaster refused and said he would go public, he says Bubeck asked, "Why would you ruin your career?" and added, "If you don't want me to be nice, then I don't have to be nice."

Buckmaster stopped short of a direct accusation. "I have not seen OpenAI's proof. I do not know what their model did, or how. I do not know whether our data was used. I am not accusing anyone of anything. I am stating what I was told, when, and what was proposed to me."

Bubeck, who led the OpenAI project, posted on X that "a series of false and inflammatory allegations against me are currently circulating on social channels," said he "came into the discussion following academic norms," and promised a fuller statement. In a press briefing he said: "We did not use their prompts or proofs to prompt our models or direct our agents. We, whether it's the researchers or the agents, did not see any of their work until they were released publicly yesterday night." He added that OpenAI recognizes "the priority of Levent Alpöge and Tristan Buckmaster's work."

OpenAI's announcement says "no specific user data was accessed in order to solve this problem," while conceding that "while unlikely, we cannot rule out that de-identified data derived from their usage of our products helped improve our models." Buckmaster responded on Mastodon that OpenAI is "openly admitting they used training data from a period after we found our result."