During development, morphogen gradients precisely determine the position of gene expression boundaries despite the inevitable presence of fluctuations. Recent experiments suggest that some morphogen gradients may be interpreted prior to reaching steady-state. Theoretical work has predicted that such systems will be more robust to embryo-to-embryo fluctuations. By analyzing two experimentally motivated models of morphogen gradient formation, we investigate the positional precision of gene expression boundaries determined by pre-steady-state morphogen gradients in the presence of embryo-to-embryo fluctuations, internal biochemical noise and variations in the timing of morphogen measurement. Morphogens that are direct transcription factors are found to be particularly sensitive to internal noise when interpreted prior to steady-state, disadvantaging early measurement, even in the presence of large embryo-to-embryo fluctuations. Morphogens interpreted by cell-surface receptors can be measured prior to steady-state without significant decrease in positional precision provided fluctuations in the timing of measurement are small. Applying our results to experiment, we predict that Bicoid, a transcription factor morphogen in Drosophila, is unlikely to be interpreted prior to reaching steady-state. We also predict that Activin in Xenopus and Nodal in zebrafish, morphogens interpreted by cell-surface receptors, can be decoded in pre-steady-state.