We infer supermassive black hole (SMBH) accretion rates and Eddington ratios as a function of SMBH/host galaxy mass and redshift with the empirical TRINITY model of dark matter halo--galaxy--SMBH connection. The galaxy--SMBH mass and growth rate connection from TRINITY is constrained by galaxy observables from \(0<z<13\) and SMBH observables from \(0<z<6.5\). Key findings include: 1) the ratio between cosmic SMBH accretion rate and galaxy star formation rate stays constant at \(\sim 2\times 10^{-3}\) from \(z=0-4\), and decreases by 2 orders of magnitude from \(z=4-10\); 2) the average SMBH Eddington ratio \(\overline{\eta}\) increases towards higher redshifts, nearly reaching \(\overline{\eta}=1\) at \(z\sim 10\); 3) at fixed redshift for \(z<3\), SMBHs/galaxies with higher masses have lower \(\overline{\eta}\), consistent with AGN downsizing; 4) the average ratio of specific SMBH accretion rate (\(\overline{\mathrm{SBHAR}}\)) to average specific star formation rate (\(\overline{\mathrm{SSFR}}\)) is nearly mass-independent, with a value \(\overline{\mathrm{SBHAR}}/\overline{\mathrm{SSFR}}\sim 1\), which decreases slightly from \(z=10\) to \(z=0\); 5) similar to galaxies, SMBHs reach their peak efficiency to convert baryons into mass when host halos reach \(10^{12} M_\odot\); 6) given galaxy and SMBH growth histories from TRINITY, the local descendants of \(1<z<11\) overmassive JWST AGNs will remain outliers from the local SMBH mass--galaxy mass relation. These findings combine to give a simple explanation for massive (\(10^9-10^{10}M_\odot\)) quasars at \(z>6\): at these redshifts, dark matter halos experience \(\sim\)Eddington specific growth rates, driving \(\sim\)Eddington specific growth rates in both galaxies and SMBHs.