Progress of the Anthropic Model
On August 12 at 14:30, a material was published about a new unreleased Anthropic model that has made progress in researching the Riemann hypothesis. This mathematical question concerning the distribution of prime numbers has remained unsolved for over 150 years. A reward of 1 million dollars has been offered for proving the Riemann hypothesis, yet as of the publication, no researcher has received this reward.
According to information provided by TechCrunch, the Anthropic model autonomously coordinated its work for about 36 hours. During this time, it examined 650 different ideas, engaged 60 AI agents, and generated approximately 31 million tokens. The model was able to increase the lower bound of the number of solutions for which the validity of the hypothesis was confirmed. The results of the work were verified by two mathematicians from Anthropic, and the proof was formalized using Lean.
The Role of Artificial Intelligence in Mathematics
In the process of working on the Riemann hypothesis, 2 agents focused on developing key mathematical ideas, 13 agents assisted in finding solutions, while other agents verified arguments or worked on preparing initial draft text for the research. This case is one of many where artificial intelligence demonstrates its usefulness in the field of mathematics. Previously, in 2026, AI had already helped solve a number of Erdős problems, and OpenAI reported on 10 mathematical results obtained by its internal Astra model. Anthropic also has separate work on disproving the Jacobian hypothesis.
Recently, there has been increasing discussion in the mathematics community about the impact of artificial intelligence on traditional principles of mathematical science. In June, a group of mathematicians signed a declaration expressing concerns that AI could alter these principles. The declaration stated that the authors of proofs should personally be responsible for their correctness. Fields Medalist Timothy Gowers noted that the impact of AI could change mathematics in both problematic and positive ways.
“The impact of AI could change mathematics in both problematic and positive ways.”
Timothy Gowers
The progress of the Anthropic model in researching the Riemann hypothesis highlights the growing role of artificial intelligence in mathematical research. This case could represent an important milestone in the development of collaboration between mathematics and technologies utilizing AI. Meanwhile, the increasing discussion of ethics and responsibility in the use of AI in mathematics may lead to the formation of new standards and principles in this field.
As the field of mathematics evolves, the role of artificial intelligence continues to spark debates regarding its influence on traditional methodologies. In a recent prediction, experts, including Elon Musk, suggest that AI could surpass human intellect within five years, prompting further considerations about the implications for mathematical research. To explore these insights and their potential impact on the future of intelligence, see more about AI's evolving capabilities.