The Riemann hypothesis is equivalent to the prime number theorem holding with a square-root-size error term.
Assessment
The claim traces to reliable primary sources through a clear chain of evidence.
The prime number theorem says that the number of primes up to x is asymptotically the logarithmic integral li(x); the Riemann hypothesis governs how far the true count can stray from that average. The precise statement is a classical theorem: the Riemann hypothesis holds if and only if π(x) − li(x) is O(√x log x), or equivalently O(x^(1/2+ε)) for every positive ε. The forward direction was proved by Helge von Koch in 1901 from Riemann's explicit formula, and Schoenfeld later made it explicit (under the hypothesis, |π(x) − li(x)| < √x log x / (8π) for x ≥ 2657); the converse follows because an error term that small forces the logarithmic derivative of the zeta function to be analytic to the right of the critical line, leaving no room for zeros there.
The equivalence is a special case of a more general correspondence: the growth exponent of the prime number theorem's error term equals the supremum of the real parts of the zeta zeros, so the location of the zeros and the size of the prime deviation are two descriptions of one quantity. Littlewood's theorem of 1914, that π(x) − li(x) changes sign infinitely often and is not O(√x log log log x / log x), shows that the square-root bound cannot be improved by much, so the hypothesis pins down the true order of the deviation rather than merely bounding it. The theorem appears in every standard text on the distribution of primes and is not disputed; it is the usual explanation, given for instance in the Clay Mathematics Institute's official problem description, of why the hypothesis is central to prime number theory.
Full reasoning: the evidence and decisions behind this verdict
The claim, in its precise form, is von Koch's theorem together with its converse. Both directions are recorded as load-bearing premises and both are textbook results. The forward direction, that the Riemann hypothesis gives π(x) − li(x) = O(√x log x), was proved by von Koch in Acta Mathematica 24 (1901): under the hypothesis each term x^ρ/ρ of the explicit formula for the Chebyshev function has modulus √x/|ρ|, and summing over zeros up to height about x gives ψ(x) − x = O(√x (log x)²), which passes to the stated bound for π(x) by partial summation; Schoenfeld (Mathematics of Computation, 1976) gives the explicit constant 1/(8π) for x ≥ 2657. The converse, that an error term O(x^(1/2+ε)) for every ε implies the hypothesis, follows from writing −ζ′(s)/ζ(s) − s/(s−1) as s times the Mellin transform of ψ(x) − x, which converges and is analytic for real part greater than 1/2 under the error bound, so ζ has no zeros there. Both appear, for example, as Theorem 30 of Ingham's The Distribution of Prime Numbers (1932) and in Davenport's Multiplicative Number Theory and Montgomery and Vaughan's Multiplicative Number Theory I.
Two further results establish that "characterizes" is the right word. The general theorem that the error exponent equals the supremum Θ of the zeros' real parts shows the correspondence is exact for any value of Θ, the hypothesis being the case Θ = 1/2. Littlewood's 1914 oscillation theorem shows unconditionally that the deviation is at least of order √x up to logarithmic factors infinitely often, so the square-root bound is essentially the best possible; Bombieri's problem description makes exactly this point. The claim presupposes the prime number theorem itself, proved in 1896 and undisputed.
All three recorded sources affirm the claim and none denies it: the Clay Institute's problem page (as a gloss), Bombieri's official problem description for the Clay Institute (www.claymath.org/wp-content/uploads/2022/05/riemann.pdf, section II, which states the equivalence with the O(√x log x) term and cites Littlewood's oscillation), and Cisło and Wolf's survey of criteria equivalent to the hypothesis (arxiv.org/abs/0808.0640), which attributes the equivalence to von Koch and also gives the O(x^(1/2+ε)) form. No source found disputes the theorem, and no mathematician has questioned it in the century since its proof. The status is therefore verified in the sense of an accepted, long-standing, independently expounded proof; no machine-checked formalization has been examined. The credence is left a hair below one only because the canonical wording ("square-root-size error term") is informal and a reader could mean by it a bound without the logarithmic factor, which is not what the theorem gives; the O(x^(1/2+ε)) reading and the O(√x log x) reading are both correct. What would change the conclusion: nothing short of an error in a proof that has been reproduced in every standard text for a century.
Decomposition
How this claim breaks down: each argument is stated as it runs, with its subclaims linked inline. ↗︎ opens a subclaim; the map shows how they fit together.
Because the Riemann hypothesis yields the error bound O(√x log x) for the prime-counting function and, conversely, an error bound O(x^(1/2+ε)) for every ε forces every zero onto the critical line, the hypothesis and the square-root-size error term each imply the other, which is the claimed equivalence.
The inference is immediate: two implications in opposite directions make an equivalence. Its weight rests entirely on von Koch's forward direction and the converse from the error bound to the hypothesis, both textbook theorems proved over a century ago and never questioned, so the argument establishes the claim outright.
Given that the growth exponent of the prime number theorem's error term equals the supremum of the real parts of the zeta zeros, the Riemann hypothesis (supremum one half) is exactly the statement that this exponent is one half; and because the deviation is unconditionally at least of order √x up to logarithmic factors, the square-root bound is essentially the smallest the deviation could have, so the hypothesis pins down the true size of the deviation rather than merely bounding it.
The inference goes through: if the error exponent equals the supremum of the zeros' real parts, then that supremum being one half is the same statement as the error term having exponent one half, and Littlewood's lower bound shows the resulting square-root order is the true order of the deviation, not merely an upper bound. The argument lives on the exponent correspondence, a standard theorem in Ingham's tract, with Littlewood's oscillation theorem supplying the sharpness; both are settled and the argument corroborates the direct equivalence rather than being needed for it.
The claims this one rests on directly, not gathered into a named line of reasoning.
- assumesbackground the parent's framing takes as givensteward instructions →The number of primes up to x is asymptotically x divided by the natural logarithm of x. ↗︎
Provenance
Where this claim has been said, linked to its canonical form.
The prime number theorem determines the average distribution of the primes. The Riemann hypothesis tells us about the deviation from the average.
Introductory description of the significance of the Riemann hypothesis.
Asserted without evidence of the source's own. The Clay Institute's page states the point as a one-line gloss, with no proof or citation; the precise theorem it condenses (von Koch's equivalence) is set out in Bombieri's official problem description, which the page links to.
The validity of the Riemann hypothesis is equivalent to saying that the deviation of the number of primes from the mean Li(x) is π(x) = Li(x) + O(√x log x); the error term cannot be improved by much, since it is known to oscillate in both directions to order at least Li(√x) log log log x (Littlewood).
Section II ("History and significance of the Riemann hypothesis") of the official Millennium Prize problem description, arguing that the hypothesis is central to the question of how well Li(x) approximates π(x).
The source's own evidence bears what it asserts. Bombieri states the equivalence in its precise form, with the O(√x log x) error term, and adds Littlewood's oscillation result to show the bound is nearly sharp. He gives no proof, but the statement is a standard theorem (von Koch, 1901) and the description is written by an authority in the field for the Clay Institute. Worth reading closely: This is the precise statement behind the Clay Institute's one-line gloss, and it also states the sharpness result; a Steward checking what the discourse means by the claim should read this passage (section II, page 4).
In this spirit in 1901, Koch proved [29] that the Riemann Hypothesis is equivalent to the following error term for the expression for the prime counting function: π(x) = Li(x) + O(√x ln(x)). Another similar criterion is RH ⇔ π(x) = Li(x) + O(x^{1/2+ε}) for each ε > 0.
Survey of criteria equivalent to the Riemann hypothesis, listing von Koch's prime-counting error term first among the classical equivalences.
How these sources relate
- https://www.claymath.org/millennium/riemann-hypothesis/ restates https://www.claymath.org/wp-content/uploads/2022/05/riemann.pdf, stating it more cautiously than that document warrants. Both documents are published by the Clay Mathematics Institute for the same prize problem; the web page is the popular summary and points readers to Bombieri's description as the official statement. The page drops the precise error term and the equivalence, stating only that the hypothesis "tells us about" the deviation, so the claim crosses in a vaguer, weaker form than the theorem Bombieri states.
Cite this claim: a formal citation with its evidence attached
Contribute
Every judgment on this page is open to challenge. A contribution is evaluated on its merits by the reviewer; if it succeeds the page changes, and if it does not, the reasons are stated. Either way the exchange becomes part of the claim’s public record.
The attention this claim received was paid for by a funded mandate. Funding buys only scheduling: it can make an assessment happen sooner, or reach deeper into a subtree. It has no influence on what the assessment concludes, and none on which claims enter the graph; assessments run under the same public standards whoever pays, funders never see or shape a verdict before anyone else, and mandates that attempt to steer conclusions are refused.
Created by extractor · Sep 13, 2026. Every judgment on this page is accompanied by a reasoning trace.