AI Proved the Quasi-Riemann Hypothesis. What Happens Now?
Unlike other recent AI proofs, this one has been unpacked by human experts. The implications, one mathematician said, are “staggering.” The post AI Proved the Quasi-Riemann Hypothesis.
On October 6, 2026, OpenAI announced 377 new results, including a claim to have proven the quasi-Riemann hypothesis, a weaker version of the famous Riemann hypothesis. This achievement is described by Jared Duker Lichtman of Stanford University as the most significant use of AI in mathematics to date. The Riemann hypothesis, posed over 160 years ago by Bernhard Riemann in 1859, is considered one of the most important unsolved problems in mathematics. Resolving it would deepen understanding of prime numbers, which are fundamental building blocks of arithmetic. The quasi-Riemann hypothesis implies many consequences of the original hypothesis, making its proof a major milestone despite being less comprehensive.
The Riemann hypothesis concerns the behavior of the Riemann zeta function (ζ(s)), a mathematical function that extends the sum 1 + 1/2^s + 1/3^s + 1/4^s + ... to complex numbers. A complex number has two parts: a real part and an imaginary part, which is a multiple of the square root of −1 (written as i). The hypothesis states that all nontrivial zeros of the zeta function—where it equals zero but is not a negative even number—lie on a vertical line where the real part of s is 1/2. This line is often called the critical line. The Riemann hypothesis suggests that primes, which thin out as numbers grow larger, follow predictable patterns governed by the zeros of this function.
Prime numbers are often called the 'atoms of arithmetic' because every integer greater than 1 can be uniquely broken down into primes. The distribution of primes is irregular but follows an average pattern described by the prime number theorem. This theorem states how quickly primes thin out as numbers increase. The Riemann hypothesis, if proven, would confirm that deviations from this average are always small. Mathematicians have numerically verified that the first 12 trillion nontrivial zeros of the zeta function lie on the critical line, but they cannot be certain this holds for all zeros. A single exception would have profound implications for the understanding of primes.
The quasi-Riemann hypothesis is a weaker version of the original hypothesis. Instead of requiring all nontrivial zeros to lie exactly on the critical line (real part = 1/2), it only requires them to be contained within a narrower vertical strip than the known strip between 0 and 1. OpenAI presented two versions of their proof. The first, dated September 30, 2026, claimed to show that all nontrivial zeros must lie between 1/8 and 7/8, and was formally verified using the Lean programming language. The second version, completed a week later on October 5, 2026, proved a weaker result: that the zeros must have a real part between 1/12 and 11/12. The second version is shorter (49 pages compared to 199 pages) and is the focus of current mathematical scrutiny.
Mathematicians, including Maksym Radziwiłł of New York University and Alexander Dunn of the Georgia Institute of Technology, are analyzing the AI-generated proofs. Radziwiłł noted that the papers are poorly written by traditional standards and lack clarity, but the 11/12 result is being prioritized due to its brevity and similar techniques. The implications of the quasi-Riemann hypothesis are described as 'staggering' by Radziwiłł, as most consequences of the full Riemann hypothesis already follow from this weaker version. The proofs rely on techniques developed by Radziwiłł and Dunn, and mathematicians are now working to understand what the results could mean for future research in number theory.
