Artificial intelligence found a new result regarding the Riemann function
Correction (22.11.2026): The previous text stated that "the Riemann hypothesis was proven" — this is incorrect. The hypothesis has not been proven; the obtained result improved the lower bound related to it. Source: anthropic.com/research/riemann-zeta
Anthropic announced that its Claude artificial intelligence, in collaboration with mathematicians, has obtained a new mathematical result on the Riemann zeta function. The Riemann hypothesis has been unproven since 1859 and remains an open problem — this result does not prove it. The obtained result improved an important metric related to the hypothesis: the proportion of zeros proven to lie on the critical line increased from 41.6 percent to over 67 percent.
The result was verified by Anthropic mathematicians and external experts Brian Conrey and Dan Goldston, and the proof was formally verified in the Lean system. The company itself emphasized that it does not expect these methods to lead to a proof of the Riemann hypothesis.