Odd one...
Low mag density
Checked it computationally and against the literature. It's the C16 (Al₂Cu / FeSn₂-type) structure, decorated with a 1:3 Mn:Fe ordering on the transition-metal sites. Not a new framework, and not a registered compound.
What the file actually is:
It parses clean (12 atoms, min distance 2.66 Å, no overlap artifacts). The metals sit on chains along c with M–M = 2.66 Å = c/2, each metal surrounded by 8 Sn at ~2.80 Å, and the Sn form the distorted square-net framework at x ≈ 0.162, the classic C16 8h parameter. Stoichiometry M₄Sn₈ = 2 MSn₂, i.e. the Sn-rich MSn₂ family, not a Heusler-like 1:1:1.
StructureMatcher against a reference C16 cell built from those coordinates matches at RMS 0.004, essentially exact. The P422 cell here is just the I4/mcm C16 conventional cell with the body-centering and glides broken by the Mn ordering.
The "odd" part: Mn occupies exactly one of the four metal-chain sites (½,½,½); the other three are Fe. So it's an ordered (Fe₀.₇₅Mn₀.₂₅)Sn₂ superstructure. The experimental literature knows this system only as a disordered solid solution: the FeSn₂–MnSn₂ C16 pseudobinary (Havinga, Damsma, Hokkeling, J. Less-Common Met. 27, 169 (1972), https://doi.org/10.1016/0022-5088(72)90008-8) and the ⁵⁷Fe Mössbauer study of Mn₁₋ₓFeₓSn₂ (Shiraishi et al., JJAP 32S3, 290 (1993), https://iopscience.iop.org/article/10.7567/JJAPS.32S3.290/pdf), which flags a magnetic-structure change right around x ≈ 0.8, i.e. near this composition. I found no ICSD, Materials Project, or published entry for an ordered MnFe₃Sn₈.
Lattice parameters sit on the end-member Vegard line: FeSn₂ a = 6.539, c = 5.325 Å (single crystal 6.533/5.320, Armbrüster et al. 2007, https://doi.org/10.1524/ncrs.2007.0033); MnSn₂ a = 6.659, c = 5.447 Å (Havinga). This file: a = 6.597, c = 5.326. c is FeSn₂'s almost exactly; a is expanded a bit toward Mn.
So if this came out of a screening or generative batch, my guess is a model reassembled the known C16 FeSn₂ framework and decorated it with a plausible 1:3 Mn/Fe ordering, or it's someone's ordered-supercell guess at the x = 0.25 composition. Either way, nothing published matches it as an ordered compound, so treat the "MnFe3Sn8" label as a hypothesis, not a phase.
As a magnet candidate it's a dud, consistent with your low-mag-density note: both end members are antiferromagnets (MnSn₂ T_N ≈ 330 K with a modulated spin structure; FeSn₂ T_N ≈ 105 K), the moment is diluted by 8 Sn per 4 transition metals, and the interesting hexagonal relatives (Mn₃Sn, Fe₃Sn₂) are a different structure family entirely.