Arithmetic formula complexity of the permanent

How large must a division-free arithmetic formula, or a valid rational formula, computing the complex n×n permanent be? Formula size here counts variable leaves in a tree, rather than gates in an unrestricted circuit.

Why it matters and what counts as progress

Why it matters: Formula lower bounds locate limitations of tree-shaped arithmetic computation without claiming the same limitation for general circuits.

What counts as progress: A stronger permanent-specific lower bound, such as order n⁴/log n, with the coefficient field, rational operations and size convention fixed.

How the problem developed

  1. 1991 · Lower bound [1]

    Noam Nisan Proved exponential formula lower bounds for the permanent in the noncommutative model; commutative arithmetic formulas remained much harder.

  2. 1 Aug 2026 · Reported claim [2][3]

    OpenAI Astra Chapter 5 reports at least n⁴/(128 log₂ n) variable leaves for division-free formulas and n⁴/(192 log₂ n) for valid rational formulas computing the n×n permanent over ℂ, for n≥32.

AI contribution

Ten Advances in Mathematics and Theoretical Computer Science

OpenAI

For the n×n permanent over the complex numbers and n≥32, Chapter 5 gives at least n⁴/(128 log₂ n) variable leaves for division-free arithmetic formulas and n⁴/(192 log₂ n) for valid rational formulas. A separate circuit lower bound has different scope.

AI involvementAI-led
ImportanceNotable
Human understandingFull
Reported

Sources

  1. Nisan, Lower bounds for non-commutative computation
  2. OpenAI, Ten Advances in Mathematics, Chapter 5
  3. OpenAI, Ten Advances in Mathematics release and attribution
  4. OpenAI, permanent formalization manifest
  5. OpenAI, permanent comparator statements
  6. Sienicki and coauthors, Chapter 5 human review
Website timeline

Publication dates and editorial changes.