OpenAI claims to solve a 100-year-old math puzzle:Solved 1 of the 7 most difficult questions in the world; answer found in 88 hours

OpenAI has claimed that its new artificial intelligence (AI) technology has solved the Navier-Stokes problem, one of the most difficult problems in the world of mathematics. It took only 88 hours to solve it. The renowned Clay Mathematics Institute selected 7 extremely difficult and unsolved mathematical problems in the year 2000. Many of these problems were over 100 years old. These were called the ‘Millennium Problems’. A prize of 1 million dollars was set for anyone who could correctly solve any one of these. Before this claim by OpenAI, only one of these seven problems had been solved. It was solved in 2003 by a Russian mathematician, Grigori Perelman. This means that if the company’s claim is correct, then two problems have been solved so far. First, understand what the Navier-Stokes problem is Suppose you spill a glass of water on the floor. The water will start spreading in all directions. Scientists use mathematics to try and predict which direction the water will go, how fast it will move, and what its flow will be like. There are certain mathematical equations for this, which are called the Navier-Stokes equations. Now, the problem for scientists was: That is exactly what the Navier-Stokes problem is. What is OpenAI’s claim? According to OpenAI, one of its new AI models has found a solution to this problem. This model has not yet been released for public use, nor has its name been disclosed. OpenAI has described its new model as being even more capable than the internal GPT-6 Astra model. GPT-6 Astra was released on September 3, 2026. However, the AI did not accomplish this task as a single system alone. OpenAI deployed approximately 10,000 AI agents to work on this task. OpenAI researcher Dan Roberts compared this to a bumblebee moving between different flowers to pollinate them. In other words, information obtained from different AI groups was transferred to other groups, which helped in finding solutions. Why are the 7 Millennium Problems so important? When an American institution selected these 7 questions in the year 2000, they were considered the most difficult unsolved problems in the world of mathematics. They were presented as major challenges to mathematics so that mathematicians around the world would try to solve them, leading to new discoveries in mathematics in the process. Before OpenAI’s claim, only one of these 7 problems had been solved. That is why OpenAI’s claim is considered so significant. What are the 7 major problems of mathematics? 1. Riemann Hypothesis 2. P vs NP 3. Navier-Stokes Problem 4. Yang-Mills and Mass Gap 5. Hodge Conjecture 6. Birch and Swinnerton-Dyer Conjecture 7. Poincaré Conjecture This was a nearly 100-year-old mathematical puzzle, proposed by French mathematician Henri Poincaré in 1904. After Perelman solved it, the Clay Mathematics Institute announced a $1 million prize for this problem in 2010. However, Perelman did not accept the prize money. How is AI able to do such difficult math? AI is given thousands of math problems to solve. It tries different methods and learns from its successes and failures. In this way, with the help of reinforcement learning, AI gradually learns methods that increase the probability of solving difficult problems. This is not the first time that AI has claimed to solve a major mathematical problem. In January, OpenAI reported solving a mathematical problem posed by the famous Hungarian mathematician Paul Erdős. Why are mathematicians concerned about AI? Many mathematicians are happy with the capabilities of AI, but some are also concerned. Terence Tao says that it is important to solve a difficult mathematical problem yourself. This helps mathematicians learn new things and increases their understanding. He explained this by giving the example of a gym. If a machine lifts 100 kg for you, the weight will be lifted, but your strength will not increase. Similarly, if AI starts solving difficult mathematical problems on its own and humans only get the answers, it is possible that humans may start losing the ability to understand mathematics and solve problems themselves. Then what is the benefit of AI? The biggest advantage of AI is its speed. A human might take a very long time to try thousands of potential paths for a mathematical problem. AI can examine a large number of possibilities in a very short time. This can help mathematicians find paths they had not thought of before. But there is another concern here. If AI finds the path itself and humans just follow that path to reach the answer, it is possible that humans might not understand the entire process of that discovery at all. Tao explained this with the example of a waterfall hidden in a forest. If someone finds the path to the waterfall through the forest themselves, they learn a lot during the entire process of finding the way. But if a machine finds the path and points it out beforehand, others can reach the waterfall directly using that same path.