We determine the non-perturbative gluon condensate of four-dimensional SU(3) gauge theory in a model-independent way. This is achieved by carefully subtracting high order perturbation theory results from non-perturbative lattice QCD determinations of the average plaquette. No indications of dimension two condensates are found. The value of the gluon condensate turns out to be of a similar size as the intrinsic ambiguity inherent to its definition. We also determine the binding energy of a B meson in the heavy quark mass limit.PACS numbers: 12.38. Gc,12.38.Bx,11.55.Hx,12.38.Cy,11.15.Bt The operator product expansion (OPE) [1] is a fundamental tool for theoretical analyses in quantum field theories. Its validity is only proven rigorously within perturbation theory, to arbitrary finite orders [2]. The use of the OPE in a non-perturbative framework was initiated by the ITEP group [3] (see also the discussion in Ref. [4]), which postulated that the OPE of a correlator could be approximated by the following series:where the expectation values of local operators O d are suppressed by inverse powers of a large external momentum Q ≫ Λ QCD , according to their dimensionality d.The Wilson coefficients C d (α) encode the physics at momentum scales larger than Q. These are well approximated by perturbative expansions in the strong coupling parameter α. The large-distance physics is described by the matrix elements O d that usually have to be determined non-perturbatively. Almost all QCD predictions of relevance to particle physics phenomenology are based on factorizations that are generalizations of the above generic OPE.For correlators where O 0 = 1, the first term of the OPE expansion is a perturbative series in α. In pure gluodynamics the first non-trivial gauge invariant local operator has dimension four. Its expectation value is the so-called non-perturbative gluon condensate