CH2NNHj has suspicious entropy in thermo libraries

This molecule

multiplicity 2
1 C u1 p0 c0 {2,S} {4,S} {5,S}
2 N u0 p1 c0 {1,S} {3,D}
3 N u0 p1 c0 {2,D} {6,S}
4 H u0 p0 c0 {1,S}
5 H u0 p0 c0 {1,S}
6 H u0 p0 c0 {3,S}

was apparently calculated apparently with CCSD(T)F12A/cc-pVTZ-F12//B3LYP/6-311++g(d,p) + BAC by Beat in the thermo_DFT_CCSDTF12_BAC library
ΔH∘f(298 K) = 79.81 kcal/mol
S∘(298 K) = 96.21 cal/(mol∗K)

and also with G4 calculation with 1D Hindered Rotors by David (automated) in the CHON_G4 library.
ΔH∘f(298 K) = 78.73 kcal/mol
S∘(298 K) = 62.97 cal/(mol∗K)

The enthalpies are remarkably close, but the entropies are very different, leading to a 50 kcal/mol disagreement in ∆G by 1500K

DB search results here.

One of them presumably has an error. What was the error and how widespread is it?

Start by comparing the CH2NNHj entries in the thermo_DFT_CCSDTF12_BAC and CHON_G4 libraries with the linked database search results. Trace how each library records the 298 K entropy, then determine which calculation is inconsistent and check whether the same discrepancy appears in other entries. Done means documenting the error and its scope.