Our evaluation of column-averaged methane (XCH₄) in models and TCCON reveals latitudinal biases between 0.4 % and 2.1 % originating from an inter-model spread in stratospheric CH₄. Substituting model stratospheric CH₄ fields by satellite data significantly reduces the large XCH₄ bias observed for one model. For other models, showing only minor biases, the impact is ambiguous; i.e., the satellite uncertainty range hinders a more accurate model evaluation needed to improve inverse modeling.