We compare radii based on Gaia parallaxes to asteroseismic scaling relation-based radii of ∼ 300 dwarfs & subgiants and ∼ 3600 first-ascent giants from the Kepler mission. Systematics due to temperature, bolometric correction, extinction, asteroseismic radius, and the spatially-correlated Gaia parallax zero-point, contribute to a 2% systematic uncertainty on the Gaia-asteroseismic radius agreement. We find that dwarf and giant scaling radii are on a parallactic scale at the −2.1% ± 0.5% (rand.) ± 2.0% (syst.) level (dwarfs) and +1.7% ± 0.3% (rand.) ± 2.0%(syst.) level (giants), supporting the accuracy and precision of scaling relations in this domain. In total, the 2% agreement that we find holds for stars spanning radii between 0.8R and 30R . We do, however, see evidence for relative errors in scaling radii between dwarfs and giants at the 4% ± 0.6% level, and find evidence of departures from simple scaling relations for radii above 30R . Asteroseismic masses for very metal-poor stars are still overestimated relative to astrophysical priors, but at a reduced level. We see no trend with metallicity in radius agreement for stars with −0.5 < [Fe/H] < +0.5. We quantify the spatially-correlated parallax errors in the Kepler field, which globally agree with the Gaia team's published covariance model. We provide Gaia radii, corrected for extinction and the Gaia parallax zero-point for our full sample of ∼ 3900 stars, including dwarfs, subgiants, and first-ascent giants.