Errors in climate models due to inadequate calculations of radiative forcing undermine researchers' ability to address important climate-related questions, including how much the atmosphere will warm as more and more carbon dioxide (CO2) is released. But, argue Brian Soden et al. in this Perspective, warming due to radiative forcing by CO2 can actually be calculated with far less uncertainty than current modeling approaches suggest. Over the past two decades, uncertainties in the radiative forcing from carbon dioxide - that is, the warming that results from the gas in the atmosphere - have kept the range of warming projected by models roughly the same. According to Soden and colleagues, CO2's role in radiative forcing is known much more accurately than climate model results suggest ; in particular, radiative forcing computed using line-by-line (LBL) calculations, which solve the radiative transfer equitation for each absorption line individually rather than averaging over absorption bands, demonstrate a far narrower range of uncertainty. These types of calculations are generally not incorporated into climate models, however, because they are too computationally expensive. Until more sophisticated methods such as these can be incorporated into climate models, model parameters should be vetted against more accurate calculations, like LBL, to reduce uncertainty in climate projections.