The infrared spectral energy distributions ( SEDs ) of z \gtrsim 5 quasars can be reproduced by combining a low-metallicity galaxy template with a standard AGN template . The host galaxy is represented by Haro 11 , a compact , moderately low metallicity , star-bursting galaxy that shares typical features of high- z galaxies . For the vast majority of z \gtrsim 5 quasars , the AGN contribution is well modeled by a standard empirical template with the contamination of star formation in the infrared subtracted . Together , these two templates can separate the contributions from the host galaxy and the AGN even in the case of limited data points , given that this model has only two free parameters . Using this method , we re-analyze 69 z \gtrsim 5 quasars with extensive Herschel observations , and derive their AGN luminosities L _ { AGN } in a range \sim ( 0.78 - 27.4 ) \times 10 ^ { 13 } L _ { \odot } , the infrared luminosities from star formation L _ { SF,IR } \sim ( < 1.5 - 25.7 ) \times 10 ^ { 12 } L _ { \odot } , and the corresponding star formation rates { SFR } \sim ( < 290 - 2650 ) M _ { \odot } / { yr } . The average infrared luminosity from star formation and the average total AGN luminosity of the z \gtrsim 5 quasar sample follows the correlation defined by quasars at z < 2.6 . We assume these quasar host galaxies maintain a constant average SFR ( \sim 620 M _ { \odot } / { yr } ) during their mass assembly and estimate the stellar mass that could form till z \sim 5 - 6 to be \langle M _ { * } \rangle \sim ( 3 - 5 ) \times 10 ^ { 11 } M _ { \odot } . Combining with the black hole ( BH ) mass measurements , this stellar mass is adequate to establish a BH-galaxy mass ratio M _ { BH } / M _ { * } at 0.1-1 % , consistent with the local relation .