← All problems

Markov Uniqueness Conjecture

For every z∈Z>0z\in\mathbb{Z}_{>0}, there is at most one pair (x,y)∈Z>02(x,y)\in\mathbb{Z}_{>0}^2 satisfying x≤y≤zx\leq y\leq z and x2+y2+z2=3xyzx^2+y^2+z^2=3xyz. Equivalently, if two positive-integer solutions of x2+y2+z2=3xyzx^2+y^2+z^2=3xyz have the same largest coordinate zz, then their two smaller coordinates agree up to exchange.

References

Progress summary

Refreshed
Open · progress reported

A new computation found no counterexample to the conjecture below an enormous bound, but the conjecture itself remains open.

Frobenius proposed in 1913 that every Markov number occurs as the largest entry of only one positive-integer solution of x2+y2+z2=3xyzx^2+y^2+z^2=3xyz, up to exchanging the smaller entries.

Known results

  • Rabideau and Schiffler proved Aigner’s fixed-numerator ordering conjecture in 2020.
  • Rabideau, Schiffler, Lee, and Li proved the fixed-denominator and fixed-sum ordering conjectures in February 2023.
  • Computation verified uniqueness for the first 18,90618{,}906 Markov numbers, up to 1014010^{140}.

October 31, 2026 computational verification

Hu Tan and Ying Zhang report an exhaustive search with no counterexample through 1050000010^{500000}, covering 239,536,376,556239{,}536{,}376{,}556 triples with distinct maximal coordinates. This is substantial finite evidence, not a proof of the conjecture, and the computation has not been independently rerun.

Current status (as of November 2026): The conjecture remains open; uniqueness is established only in partial cases and the reported computation is unverified finite evidence.

Sources

Solutions 0

No solutions have been posted yet.