The facts
The 49-page manuscript dated 5 October claims that every Dirichlet L-function, including the Riemann zeta function, has no zeros with real part greater than 11/12. Its argument uses mean-square estimates, large sieves and metaplectic theta-series techniques. [1]
A separate 199-page manuscript dated 30 September claims the stronger boundary 7/8. Quanta reports that this version was formally verified in Lean, while the shorter 11/12 version had not been formally verified at the time of its report. [2][3]
Quanta interviewed specialists including Maksym Radziwiłł and Alexander Dunn. It reports human efforts to clarify the argument and substantial criticism of the manuscripts’ exposition; formal verification and expert mathematical understanding are distinct checks. [3]
Relevant context
The full Riemann hypothesis would place all nontrivial zeta zeros on the line with real part 1/2. Both manuscripts concern weaker zero-free regions; the full conjecture remains open. [1][2][3]
Still unconfirmed
- Journal acceptance and subsequent complete checking of both manuscripts have not been established in the checked sources.
Sources
References support the numbered facts above. Source checks: 29 Oct 2026, 02:05 UTC.
Catch-up coverage of the 19 October expert assessment. No claim is made that the full Riemann hypothesis has been proved.
- The Quasi-Riemann Hypothesis, 5 October manuscript
OpenAI · Primary source · 5 Oct 2026
Original PDF retrieved; theorem, scope, prior work and argument outline read. This is a preprint, not a verified journal-publication claim.
- A Zero-Free Half-Plane Re(s) > 7/8
OpenAI · Primary source · 30 Sept 2026
Original PDF retrieved; abstract and introduction read. Same institution as source 1, not independent corroboration.
- AI Proved the Quasi-Riemann Hypothesis. What Happens Now?
Quanta Magazine · Reporting · 19 Oct 2026
Independent specialist reporting with named mathematicians. Its formal-verification and expert-assessment descriptions are attributed, not a claim that this newsroom audited the proof.