The 2018 Particle Data Group gives a value of $880.2\pm 1.0$ s for free neutron lifetime, as an average of the seven best measurements.

As can be seen, the measurements have non-overlapping confidence intervals. As discussed in (Wietfeldt 2014) the different experimental methods do not agree on the value.
Wietfeldt gives the formula of the lifetime as $$\tau_n = \left(\frac{2\pi^4\hbar^7}{m_e^5c^4f_R}\right)\frac{1}{G_V^2+3G_A^2}.$$ $f_R$ is a phase space factor for the final state and radiative corrections, $G_V$ and $G_A$ are the nucleon vector and axial vector coupling constants. $G_V=G_FV_{ud}$ where $G_F$ is a universal weak coupling constant and $V_{ud}$ is the first element of the CKM matrix. Good measurements of $\tau_n$ would help determine the values of these constants better; current estimates of $V_{ud}$ are from nuclear or pion decays.
One can do theoretical calculations using weak force diagrams, but I get the feeling that the numerical value still depend on a lot of empirically measured constants.