We propose the option realized variance as a new observable covariate that integrates high frequency option prices in the inference of option pricing models. Using several simulation and empirical studies, this paper documents the incremental information offered by this variable. Our empirical results show that the information contained in this new covariate improves the inference of model variables such as the instantaneous variance, return jumps, and variance jumps of the S&P 500 index. Parameter estimates indicate that the risk premium breakdown between jump and diffusive risks is affected by the omission of this information.