Aims. In this study I assess how existing data for the solar analogues 16 Cyg A and B, in particular the asteroseismic measurements obtained from Kepler, constrain theoretical stellar models. The goal is two-fold: first to use these stars as benchmarks to discuss which precisions can realistically be expected on the inferred stellar quantities; and second to determine how well “non-standard” prescriptions, such as microscopic diffusion and overshoot, are constrained. Methods. I used a Bayesian statistical model to infer the values of the stellar parameters of 16 Cyg A and B. I sampled the posterior density of the stellar parameters via a Markov chain Monte Carlo algorithm, tested different physical prescription, and examined the impact of using different seismic diagnostics. Results. General good agreement is found with several recent modelling studies on these stars, even though some discrepancies subsist regarding the precise estimates of the uncertainties on the parameters. An age of 6.88 ± 0.12 Gyr is estimated for the binary system. The inferred masses, 1.07 ± 0.02 M_⊙ for Cyg A and 1.05 ± 0.02 M_⊙ for Cyg B, are shown to be stable with respect to changes in the physical prescriptions considered for the modelling. For both stars, microscopic diffusion has a significant effect on the estimates of the initial metallicity. Overshoot is confined to very small regions below the convective zone. I show that a proper treatment of the seismic constraints is necessary to avoid biases in the estimate of the mass.