← All papers
First page of An upper bound of 61 for the Ramsey number $R(3,3,3,3)$

An upper bound of 61 for the Ramsey number $R(3,3,3,3)$

Milos Tatarevic

math.CO Oct 27, 2026 · v1
Lean 4 formalizes the combinatorial reductions to 56,723 SAT formulas, checks the template-classification certificates, and emits the CNF formulas. The SAT refutations themselves are an unverified premise.
We improve the upper bound on the classical Ramsey number $R(3,3,3,3)$ from 62 to 61. The proof is based on a global incidence count, classifications of small neighborhoods and exhaustive SAT computations. We formalize the mathematical reductions in Lean and refute every resulting formula with two SAT solvers.

The best known upper bound for the classical Ramsey number R(3,3,3,3) was 62. The goal is to show that every four-coloring of the edges of K_61 has a monochromatic triangle, which lowers the bound to 61.

Assume a good four-coloring of K_61 exists. Each vertex has one of five degree profiles, and a global double count of 'tickets' (sent along critical neighborhoods) narrows the remaining cases. The proof also uses the known classification of good three-colorings of K_15 and K_16 into the U and T templates. The remaining cases (marked-neighborhood placements, C entrances, perfect balance) reduce to 56,723 CNF formulas. A Lean 4 formalization proves the reductions, checks 55 CNF/LRAT classification certificates, and emits the formulas; Kissat and CaDiCaL refute each formula.

Every formula was found UNSAT by both solvers, using about 677 CPU hours, which gives R(3,3,3,3) ≤ 61. The final Lean theorem, UniformK61.WholeProof.ramsey61, has the unsatisfiability of the 56,723 formulas as its only premise and otherwise uses only the standard axioms.

Case familyFormulasKissatCaDiCaL
A3600.311.71
B5,0967.4541.21
C entrance3460.261.42
Perfect balance30.020.03
Total56,723184.06492.71
SAT runs, CPU hours (selected families)