Will the Riemann Hypothesis be resolved before 2027?

- [2026-10-10, current] The Clay Mathematics Institute's official Riemann Hypothesis page still lists the problem's status as "Unsolved" , with no indication of any accepted resolution. https://www.claymath.org/millennium/riemann-hypothesis/ - [2026-10-05/07] OpenAI published a large batch of AI-generated manuscripts; coverage describes OpenAI publishing "722 manuscripts of mathematics produced by an internal model inaccessible to the public, including a claimed advance towards Riemann's hypothesis" , but this is explicitly a "Quasi-Riemann Hypothesis" result, not the Millennium Problem itself. https://ground.news/article/openai-solves-quasi-riemann-hypothesis-understand-what-it-means - [2026-10-07] The specific OpenAI claim is a zero-free region result for Dirichlet L-functions: "every Dirichlet L-function, including Riemann's famous zeta function, has no zeros in the half-plane where the real part of s is greater than 11/12" — a partial/weaker statement, not a proof of RH (which requires no zeros off Re(s)=1/2). https://officechai.com/ai/quasi-riemann-hypothesis-explanation/ - [2026-10-07] A detailed explainer directly addresses the question and concludes: "OpenAI did not announce a proof of the Riemann hypothesis. It released a large collection of AI-generated mathematical manuscripts, including work that it says makes progress toward famous open problems." https://en.kingofgeek.com/2026/10/openai-math-release-riemann-hypothesis/ - [2026-10-07] Same source states plainly: "the Riemann hypothesis remains unsolved unless and until a complete proof survives expert review." https://en.kingofgeek.com/2026/10/openai-math-release-riemann-hypothesis/ - [2026-10-05] Separately, Anthropic reported an unreleased Claude research model improving a numeric bound tied to RH, but "The company did not claim to have solved the Riemann hypothesis, and it was explicit about that limit." https://tech-insider.org/anthropic-claude-riemann-hypothesis-breakthrough-2026/ - [2026-10-07] Verification rigor of the OpenAI batch is contested: "about 300 of 719, according to Implicator.ai's reporting as of October 7 — that had been run through the Lean proof assistant and received a computer-checked formal verification by the time the material went public" , i.e. the majority of claims (including presumably the RH-adjacent one) lack formal verification. https://tech-insider.org/openai-math-papers-lean-verification-42-percent-2026/ - [2026-10-09] Mathematician backlash is ongoing; reviewers including Timothy Gowers and Edward Witten were involved, and "some of the proofs are reportedly so unusual in structure and style that working mathematicians, including specialists in the exact subfield, say they can't fully verify or understand them without AI assistance." This signals that even unrelated results in the batch are far from community acceptance, let alone peer-reviewed publication. https://www.mindstudio.ai/blog/openai-math-proofs-mathematicians-backlash - [2026-10-07] A finance/crypto outlet sensationally claimed "Quasi-Riemann Hypothesis Proven" and that "AI breakthrough of nuclear proportion: The Riemann Hypothesis has been solved and formally verified" — but this headline conflates the narrower "quasi" result with the actual Millennium Problem and is not corroborated by mathematical consensus or Clay Institute statements. https://www.kucoin.com/news/flash/openai-solves-722-math-problems-quasi-riemann-hypothesis-proven - [2026-02-05] Earlier in 2026, Fields Medalist Alain Connes published a survey titled "The Riemann Hypothesis: Past, Present and a Letter Through Time," which is described as "a survey of the Riemann Hypothesis, provides a comprehensive overview of 165 years of mathematical approaches to this fundamental problem... followed by an extensive survey of mathematical theories developed in pursuit of RH" rather than a claimed proof. https://arxiv.org/pdf/2602.04022 - [2026-01-19 to present] Numerous amateur/crank preprints claiming proofs or disproofs of RH continue to circulate (e.g., a Zenodo preprint submitted to Duke Mathematical Journal and "under review" as of its January 2026 posting, and a perpetually-revised arXiv "Simple Proof" updated as recently as 2026-10-07), none of which have been accepted by any peer-reviewed journal or recognized by an authoritative body. https://zenodo.org/records/18323203 ; https://arxiv.org/abs/2209.01890 BOTTOM LINE: As of 2026-10-10 the Riemann Hypothesis in its official Clay Mathematics Institute form remains explicitly listed as "Unsolved," and the most prominent recent development (OpenAI's early-October 2026 "Quasi-Riemann Hypothesis" release) is a partial/weaker result that commentators and the company itself distinguish from a proof of RH; no peer-reviewed proof, authoritative-body recognition, or qualifying dual-source media resolution exists, making a Yes resolution before the Jan 1, 2027 deadline currently unsupported by any primary evidence. NOT FOUND: No statement from the Clay Mathematics Institute itself addressing the OpenAI or Anthropic 2026 announcements; no evidence of any peer-reviewed journal publication of a complete RH proof or disproof; no evidence of a specific CMI prize-committee review process having been initiated for any 2026 claim.

- [2000-05-24] The Riemann Hypothesis is one of seven Clay Mathematics Institute Millennium Prize Problems, officially designated at the Millennium Meeting held on May 24, 2000, alongside the Birch and Swinnerton-Dyer conjecture, Hodge conjecture, Navier–Stokes, P vs NP, Riemann hypothesis, and Yang–Mills. https://graphsearch.epfl.ch/en/concept/193053/concepts - [2010-01-01] In the 26-year history of the Millennium Prize Problems, the only Millennium Prize problem solved to date is the Poincaré conjecture, with the Clay Institute awarding the monetary prize to Grigori Perelman in 2010 — i.e. only 1 of 7 (~14%) has ever been resolved, and that process (proof posted 2002–2003, formal verification and prize award 2010) took roughly 7-8 years from preprint to official confirmation, illustrating how long authoritative verification takes even for a correct proof. - [2018-09-24] A high-profile false-positive precedent exists: Sir Michael Atiyah claimed a RH proof in 2018, but mathematicians immediately expressed skepticism, calling it "too vague and unspecific" to be a real proof , and the claim was never accepted — showing that announced "proofs" historically do not resolve the conjecture. - [2026-08-31] A 2026 episode shows the volatility/noise risk in this reference class: the market's primary move was driven by research from an AI model (Claude) that increased the known lower bound for zeros on the critical line to ~67.2%, after which the price moderated as a Scientific American article clarified the AI's work was an incremental advance, not a complete solution . https://www.octagonai.co/markets/science-and-technology/physics-math/when-will-the-riemann-hypothesis-be-solved - [2026-09-09] Despite that AI-driven spike, outside-view forecasts remain far out: Metaculus forecasters estimate a median resolution date of April 2044 , and the market prices RH as likely to be solved before 2035 at 56%, while a model-based estimate put this probability at 37.6%, with experts calling a near-term AI solution unlikely . https://www.octagonai.co/markets/science-and-technology/physics-math/when-will-the-riemann-hypothesis-be-solved - [2026-09-09] Octagon's own "Before 2028" estimate for the broader solved-by date market stood at market 20% vs. model 11.7%, with the stated rationale: "Experts clarified that recent AI advancements were misinterpreted and do not offer a near-term solution." https://www.octagonai.co/markets/science-and-technology/physics-math/when-will-the-riemann-hypothesis-be-solved - [2024 (snapshot)] A separate long-running Metaculus question "When will the Riemann Hypothesis be resolved?" (68 forecasters) gives a current community estimate of May 2035, with a 90% credible interval spanning 06 May 2028 to Dec 2059 — the lower bound of that interval (2028) is already after the Kalshi contract's Jan 2027 deadline. https://www.metaculus.com/questions/6494/when-will-the-r-hypothesis-be-resolved/ - [2021-2023 snapshot] A related Metaculus series on Millennium Prize Problems in general found community medians of roughly 10 Sep 2030 for "next Millennium Prize Problem solved" (closing Jan 2030) and May 2063 specifically for the Riemann hypothesis , and a newer version of that question put the Riemann-specific median at Jun 2046 . https://www.metaculus.com/questions/15838/millennium-prize-problems-solution-date/ - [~2023 snapshot] On Metaculus, conditional on the question "if RH is proved by 2100, will it be proved true (not false)?", forecasters gave 92% chance — relevant only to direction of outcome, not timing, but confirms virtually no credible-mathematician base rate exists for a near-term disproof/counterexample either. - [2026-04 (reported)] Expert sentiment: mathematician Peter Sarnak expressed an expectation in April 2026 that he would live to see the resolution of at least one major Millennium Prize problem, including the Riemann Hypothesis, which experts consider a potential scientific "earthquake" — implying experts see resolution as plausible within decades, not within the next year. - [general/historical] Riemann wrote down the conjecture in 1859; it remains unproved 165 years later , and incremental partial results (Hardy 1914, Levinson 1974, and modern zero-density/zero-proportion improvements through 2026) have steadily raised bounds without resolving the full conjecture, reflecting a base rate of slow, incremental — not sudden — progress. BOTTOM LINE: Across every available reference point — the 26-year track record of Millennium Problems (1 of 7 solved, over a ~7-8 year proof-to-confirmation lag), repeated false-alarm "proof" claims (Atiyah 2018, the 2026 AI-bound episode), and aggregated forecaster/market medians clustering in the 2035–2046+ range with 90% intervals not even starting until 2028 — the base rate strongly implies resolution before January 2027 would be a historically unprecedented, out-of-distribution event. NOT FOUND: No historical instance of a Millennium Prize problem (or RH specifically) being resolved on a sub-1-year notice/surprise basis; no explicit long-run "base rate per year" statistic published by CMI or AMS; current (October 2026) up-to-date Metaculus/Manifold point estimates beyond the September 2026 snapshots cited above.

YES means: the Riemann Hypothesis, in its official Clay Mathematics Institute formulation, gets a complete proof/disproof published in a qualifying peer-reviewed venue, or is recognized as resolved by an authoritative body, or is reported as resolved by two Source Agencies without subsequent retraction — all before Jan 1, 2027, and that status holds through market expiration. - [2026-09-11] The Clay Mathematics Institute already showed it can engage with an AI-generated Millennium Problem claim within days rather than years: CMI's own statement said it "the Navier-Stokes problem has apparently been settled" and hopes "to see waves of new human understanding unleashed as the innovations behind this work are analyzed and interrogated." This came just three days after OpenAI's internal claim, establishing a precedent that CMI will publicly acknowledge a serious AI-derived result almost immediately rather than staying silent for years. - [2026-09-21] OpenAI's Navier-Stokes claim arrived alongside creation of a standing "independent advisory group" of mathematicians, the same infrastructure (AGMAI) now vetting the October RH-adjacent release — meaning a review pipeline capable of fast-tracking scrutiny of a future full-RH claim already exists and is operational. - [2026-10-06] OpenAI's 722-manuscript release explicitly includes "results bearing on three of the Millennium Prize Problems still unsolved, among them the Riemann hypothesis" , and a Rutgers mathematician reacted to one RH-related proof saying "If a human did this, it would be an instant Fields Medal, no questions asked" — signaling that at least some specialists view the underlying technique as unusually strong, not crankish. - [2026-10-07] The zero-free region claim was tightened over the same week from 7/8 to 11/12: OpenAI's paper claims that "there are no zeros anywhere to the right of the line at 11/12, roughly 0.917" , showing the gap to the full hypothesis's 1/2 line was already narrowing within days, not years, of the first announcement — a visible trajectory a forecaster could extrapolate toward further compression before January 2027. - [2026-10-07] Separately, the same release "shows that neither the zeta function nor many of its generalizations, called 'Dirichlet L-functions,' have a 'Siegel zero,' a certain, especially problematic type of zero that would contradict the hypothesis" — eliminating one of the few remaining classical obstructions that number theorists have long treated as a major roadblock to progress toward RH. - [2026-09 (reported)] An OpenAI/Anthropic-adjacent assessment noted one mathematician's view that "One proves the 'quasi-Riemann hypothesis,' which is a huge step toward the holy grail of proving the full thing," with "Plenty of applications of the Riemann hypothesis did not need the full Riemann hypothesis" — reflecting genuine (if contested) expert sentiment that this line of AI-driven work could plausibly be "progress," not noise, toward the full conjecture. - [2026-10-06] OpenAI's compute-per-result figures show the model needed only "the average result required the equivalent of roughly three hours of ChatGPT Pro thinking" for most of the 372 result families, illustrating the sheer throughput of automated proof search now available; if even a small fraction of similarly-scaled attempts target the remaining gap, the compute cost of further attempts before January 2027 is trivial compared to historical human effort. - [2026-10-07] Lean formalization coverage, while partial, is substantial and growing: OpenAI posted 722 manuscripts in 372 result families from about 4,000 problems, and "links Lean proofs for 235 of the 372 families" , with the quasi-RH entry itself having "a Lean proof" — giving the resolution criterion's bar of "peer-reviewed" or "authoritative body" recognition a faster technical on-ramp than past eras, since formal verification can substitute for some traditional peer-review delay. - [2026-10-06] The release's own framing by its authors was explicit that this is a stepping stone rather than a terminus — Hector Pasten noted "it's too early to say" whether this is progress toward the full hypothesis; "Some people claim it is progress toward Riemann; other people say it's not. Usually the word progress makes sense after the fact, when you see how it eventually gets solved" — which, while cautious, keeps open a live, expert-endorsed possibility channel that the current trajectory culminates in a full proof. - [2026-10-09] Despite verification concerns, the sheer organizational weight now mobilized — Fields medalists, AGMAI, Lean formalization infrastructure, multiple competing frontier labs (OpenAI, Anthropic, DeepMind) racing on overlapping RH-adjacent problems — represents a step-change in the number of credible, well-resourced attempts occurring simultaneously within a single quarter, raising the tail probability of a qualifying breakthrough materializing before year-end relative to any prior period in RH's 165-year history. BOTTOM LINE: The strongest case for YES rests on the unprecedented acceleration of AI-driven number theory in September–October 2026 — a fast-moving CMI/AGMAI review apparatus, a demonstrated CMI willingness to publicly acknowledge "apparently settled" status within days (Navier-Stokes), and a zero-free-region bound that tightened from 7/8 to 11/12 within the same week — meaning the pathway to YES is not implausible on its face, even though no current claim actually reaches the full critical-line statement. NOT FOUND: No evidence of any lab or individual claiming to have closed the gap between the zero-free region (currently ~11/12) and the full critical line (1/2); no CMI statement addressing the Riemann Hypothesis specifically (only Navier-Stokes); no timeline or roadmap published by OpenAI, Anthropic, or DeepMind for attempting the complete Riemann Hypothesis before January 2027.

YES means: the Riemann Hypothesis (official CMI formulation) is credibly established as proved, disproved, or found to have a counterexample via peer-reviewed publication, authoritative-body recognition, or dual-sourced media confirmation (per the contract's three Payout Criterion pathways), with no retraction before Jan 1, 2027 — i.e., a Yes outcome requires a decisively confirmed resolution in less than three months from today. - [2018-09-26] The Clay Mathematics Institute's own binding rules require that, after publication in a Qualifying Outlet, a proposed solution must survive "rigorous examination" by the global mathematics community for a minimum of two (2) years before CMI will even decide whether the solution "merits detailed consideration" — meaning even a hypothetically perfect proof submitted today could not complete CMI's own institutional recognition pathway (criterion B of the contract) before the Jan 2027 deadline. https://www.claymath.org/wp-content/uploads/2022/03/millennium_prize_rules_0.pdf - [2018-09-26] The rules further specify that CMI's first evaluation stage alone requires the proposed solution to achieve general acceptance "as determined by CMI in its sole discretion," with CMI explicitly stating "No CMI-affiliated entity will accept invitations, requests or demands to recognize the status of a Proposed Solution" — ruling out any shortcut acknowledgment under outside pressure or media hype. https://www.claymath.org/wp-content/uploads/2022/03/millennium_prize_rules_0.pdf - [2026-09-11] A directly analogous 2026 precedent exists within the same Millennium Prize framework: after OpenAI's Sept. 8, 2026 announcement on Navier–Stokes, CMI publicly called the problem "apparently" settled and stated that evaluation and assignment of credit would follow its prize rules , yet the announcement did not award a prize, and CMI continues to label the problem "Active" as of the Sept. 11 statement — demonstrating that even a headline-grabbing, seemingly strong claimed resolution to a sibling Millennium Problem did not translate into any formal "resolved" status within weeks, reinforcing that the institutional bar for official recognition is slow and conservative by design. https://mathworld.wolfram.com/MillenniumPrizeProblems.html - [2026-09-12] As of this Wolfram MathWorld snapshot (updated Sept. 12, 2026), the Riemann Hypothesis itself is still flatly listed as "unsolved" in CMI's own classification table, with only the Poincaré conjecture marked "solved" in the 26-year history of the program and Navier–Stokes merely "active" despite the OpenAI claim — underscoring that no credible claim specific to RH has even reached the "apparently settled" status Navier-Stokes briefly achieved. https://mathworld.wolfram.com/MillenniumPrizeProblems.html - [2026-10-07] Even setting aside CMI's formal prize process, the contract's alternative "two-Source-Agency media" pathway (criterion C) requires reporting that a "specific publicly available written work" resolves RH with no outlet reporting an identified, uncorrected error — a bar the OpenAI release fails, since reporting as of Oct. 7 explicitly frames the batch's RH-adjacent claim as a narrower "Quasi-Riemann Hypothesis" result distinct from the Millennium Problem, and independent verification coverage notes roughly 300 of 719 manuscripts had received Lean formal verification by Oct. 7, leaving the majority (plausibly including this claim) unverified. https://tech-insider.org/openai-math-papers-lean-verification-42-percent-2026/ - [2021 (report dated June 2021)] Historical status-quo inertia is illustrated by the Kumar Eswaran RH-proof saga: an expert committee convened in January 2020 specifically because, despite the proof circulating online for roughly five years and being sent to journals, "there was a reluctance on the part of the Editors of International Journals to put the paper though a detailed Peer review" — showing that even organized advocacy and committee review over multiple years has failed to secure a qualifying peer-reviewed publication for any RH proof claim to date. https://sreenidhi.edu.in/wp-content/uploads/2022/11/A_Report_on_Riemann_Hypothesis_updated25-06-21.pdf - [2018-09-26] CMI's rules also make clear that completeness and acceptance determinations lie entirely in "the sole discretion of CMI" , and that all CMI decision-making is confidential for 50 years absent waiver — meaning even if the mathematics community informally converged on a verdict, there is no guaranteed mechanism for CMI to publicly confirm resolution within the contract window, independent of mathematical merit. https://www.claymath.org/wp-content/uploads/2022/03/millennium_prize_rules_0.pdf - [2026-10-07] Continued circulation of uncorroborated amateur/crank "proof" claims (e.g., ongoing arXiv revisions and Zenodo preprints "under review") without any journal acceptance illustrates the persistent gap between claimed solutions and qualifying publication; none of these meet the Qualifying-Outlet bar (editorial board, named referees, MathSciNet indexing) CMI's rules impose. https://www.claymath.org/wp-content/uploads/2022/03/millennium_prize_rules_0.pdf BOTTOM LINE: Both the letter of CMI's own Millennium Prize rules (mandating a minimum two-year post-publication waiting period before any official recognition) and the closely analogous September 2026 Navier–Stokes episode (an "apparently settled" claim that still left the problem labeled "Active" weeks later) show that even a strong, well-publicized claimed solution to a sibling Millennium Problem does not translate into swift official or media-confirmed resolution — making a dual-sourced, uncorrected, qualifying resolution of RH specifically before Jan 1, 2027 structurally and historically implausible. NOT FOUND: No evidence that CMI has ever relaxed or waived its two-year waiting period for any Problem; no statement from CMI specifically addressing RH in relation to the Oct. 2026 OpenAI/Anthropic announcements; no record of any journal having accepted a complete RH proof as of this writing.