I came across this rather nice model; it's firmly rooted in the SDE but uses only datasheet inputs and requires no tedious iteration to calculate the MPP. Here's a quick comparison of its error (treating the CEC SDM, with consistent inputs, as "truth"), for two modules (c-Si in first row, CdTe in second). Dotted lines in the first column show the same for PVWatts, for reference.

The comparison for those two modules makes the model seem quite promising to me.

Easy to implement; one simple lambertw call, one np.log, and the rest just elementary operations.

Reference:

E. I. Batzelis, "Simple PV Performance Equations Theoretically Well Founded on the Single-Diode Model," in IEEE Journal of Photovoltaics, vol. 7, no. 5, pp. 1400-1409, Sept. 2017, https://doi.org/10.1109/JPHOTOV.2017.2711431