← All problems

Erdős problem 647

Determine whether there exists an integer n>24n>24 such that max⁡1≤m<n(m+τ(m))≤n+2\max_{1\le m<n}\bigl(m+\tau(m)\bigr)\le n+2, where τ(m)\tau(m) denotes the number of positive divisors of mm. Equivalently, the conjecture asserts that no such nn exists.

Equivalent formulations
  1. Define M(n,k)=lcm⁡(n+1,n+2,…,n+k)M(n,k)=\operatorname{lcm}(n+1,n+2,\ldots,n+k). Determine whether there exist integers m,n,km,n,k with m≥n+km\ge n+k such that M(n,k)=M(m,k)M(n,k)=M(m,k).

    source: Background supplied with the source

References

Progress summary

Refreshed
Open · progress reported

A new computer-checked certificate verifies only a finite range of cases, so the conjecture remains open.

Erdős problem 647 asks whether the least common multiples of two separated blocks of consecutive integers can ever coincide: M(n,k)=M(m,k)M(n,k)=M(m,k) with m≥n+km\ge n+k. Erdős posed the problem in the context of his 1979 work; the main entry still lists it as open.

Known results

  • Thue–Siegel implies that, for fixed kk, only finitely many pairs (m,n)(m,n) satisfy the equality.
  • Erdős recorded the known examples M(4,3)=M(13,2)M(4,3)=M(13,2) and M(3,4)=M(19,2)M(3,4)=M(19,2).
  • A related 2025 paper reports that the equivalent divisor-function formulation remains beyond its methods.

August 2026 kernel certificate

It remains far below the prior non-kernel computational frontier and does not settle the conjecture. A January 2026 record also labels an AI-involving proof attempt incorrect.

Current status (as of August 2026): A finite exclusion range is kernel-verified, but the full conjecture and any possible exceptional cases remain open.

Sources

Solutions 0

No solutions have been posted yet.