Stachys affinis (Chinese artichoke) tubers contain 50–80% stachyose by dry weight, the most concentrated dietary source of raffinose-family oligosaccharides (RFOs) known. Because humans lack sufficient α-galactosidase activity, stachyose transits intact to the colon where microbial fermentation yields short-chain fatty acids (SCFAs). However, the quantitative impact of stachyose-derived SCFAs on host hepatic energy metabolism has not been systematically explored using genome-scale metabolic models. Three stachyose dose scenarios (Low/Mid/High: ∼25, 50, 100 g fresh tubers) were translated to SCFA availability vectors. Hepatic metabolic responses were simulated using Recon3D (10,600 reactions) and Human-GEM (13,417 reactions) under strict hepatocyte-like media, maximizing ATP maintenance flux (ATPM). FVA across multiple optimality thresholds (90–100%) and pFBA confirmed solution robustness. One-at-a-time sensitivity analysis characterized ATPM responses to individual parameter perturbations, and a ratio sensitivity sweep across six alternative SCFA profiles assessed dependence on assumed fermentation ratios. A targeted rescue experiment addressed model-specific propionate catabolism gaps. Both models showed dose-dependent ATPM increases (Recon3D: +71 to +286%; Human-GEM: +103 to +413% above baseline), with the 19–33% inter-model gap attributable entirely to Human-GEM’s functional propionate catabolism pathway. FVA confirmed near-unique optimal solutions (ATPM ranges ∼1% at 99% optimality, widening to ∼10% at 90%). Parsimonious FBA preserved identical ATPM values while reducing total flux by ∼4–14%, confirming objective robustness. SCFA ratio sensitivity across six alternative profiles showed 27– 28% ATPM variation, indicating qualitative robustness. Butyrate yielded the highest ATP per mole (∼22) in both models; propionate sensitivity was zero in Recon3D but ∼15.25 mmol ATPM/mmol propionate in Human-GEM. Reopening propionyl-CoA carboxylase (PPCOACm) in Recon3D under strict constraints converged ATPM to within 0.3–0.7% of Human-GEM, cross-validating both reconstructions. This reproducible dual-model pipeline identifies model-specific pathway gaps and provides cross-validated predictions to guide future experimental studies of how dietary SCFAs influence hepatic ATP metabolism.