ChatGPT solved a 200-year-old equation, but a mathematician accuses OpenAI of plagiarism

OpenAI claims that an internal artificial intelligence, more powerful than the GPT-6 Astra model that powers ChatGPT, solved a mathematical problem posed in the early 19th century: the Navier-Stokes equation.The equation first derives its name from French engineer Henri Navier, who posed the problem in 1822, and then from British physicist George Gabriel Stokes, who gave it its definitive form in 1845. The equation describes how fluids, such as water and air, move over time, which is very useful for predicting the weather or designing an aircraft wing.

We’re sharing a solution to the Navier-Stokes Millennium Prize Problem, one of the deepest problems at the frontier of mathematics.

The proof was produced by a group of agents, using an OpenAI next-generation model significantly more capable than GPT-6 Astra.

The problem… pic.twitter.com/8zol3BPTL4

— OpenAI (@OpenAI) September 8, 2026

Unfortunately, the equation remained a mystery since its inception. No one has ever succeeded in demonstrating how the equation works. In 2000, the Clay Mathematics Institute listed this problem among its seven “Millennium Prize Problems”. The American foundation promised one million dollars to whoever would manage to solve the mystery.

Also read: A true science-fiction film – OpenAI finally reveals how ChatGPT launched its first cyberattack

An equation solved in 88 hours

According to OpenAI, one of ChatGPT’s internal versions solved the equation in 88 hours. To do so, more than 10,000 AI agents worked together. The start-up says the entire operation cost “millions of dollars” in computing power.Mark Chen, the head of research at the company, believes that this is “a significant step in AI research”. The mathematical proof itself was then verified by GPT-6 Astra in 17 hours.

This announcement forms part of a series of mathematical breakthroughs claimed by AI giants. In May, OpenAI said it had already solved an 80-year-old enigma. More recently, the company revealed that Astra had solved ten mathematical problems that had remained unsolved for decades.

Plagiarism accusations

OpenAI’s claims were quickly challenged by Australian mathematician Tristan Buckmaster, a professor at New York University. The scientist questioned how OpenAI had arrived at its solution. He revealed that he had been working on a solution to the Navier-Stokes problem with the help of Levent Alpöge, an employee of Anthropic, OpenAI’s rival company. According to Buckmaster, OpenAI’s model followed the line of inquiry they were exploring, which differed from the approach taken by the rest of the scientific community.

The pair of researchers had used Codex, OpenAI’s AI dedicated to programming, to advance their work. Buckmaster therefore asked OpenAI whether this data had been used internally to train or guide the model… which would explain why ChatGPT ultimately managed to solve the enigma. The researcher’s theory is that the AI is in fact relying on his own research.

We congratulate Levent Alpöge and Tristan Buckmaster on their remarkable mathematical work.

We (the researchers and the agents) did not see any of their work through any means until they released it publicly — in particular, no specific user data was accessed in order to solve… https://t.co/otJKRHnQgb

— OpenAI (@OpenAI) September 8, 2026

OpenAI categorically denies having exploited the duo’s research. OpenAI researcher Sébastien Bubeck assures that he has never “used their queries or their results to guide our models or our agents”. OpenAI nevertheless admitted that it launched its research after hearing that two mathematical problems were about to be solved by other teams. Above all, OpenAI acknowledges that it cannot guarantee 100% that Buckmaster and Alpöge’s research, conducted with the help of its own AI tools, did not indirectly help improve its models.

👉🏻 Follow the latest tech news in real time: add 01net to your sources on Google, and subscribe to our WhatsApp channel.