The mid-infrared (mid-IR) spectrum in galaxies has a long history as a valuable proxy for the dust-obscured component of the star formation rate (SFR) in massive galaxies. In this work, we exploit the capabilities of the James Webb Space Telescopes (JWST) to expand our understanding of the mid-IR and its use in measuring SFRs, covering four orders of magnitude in total infrared luminosity (9≲ log L_(rm IR)/L_⊙≲13). First, using a Main Sequence sample at 1<z<1.75 from SMILES, we calibrate mid-IR-based SFRs against the P_(rm aα) emission line - a gold standard SFR indicator - from the FRESCO survey. We find that the rest-frame ∼6-8 μm emission - dominated by PAHs and probed by the Mid-Infrared Instrument (MIRI) at z∼1.3 - has a superlinear relation with SFR_(rm P_(aα)) below ∼8 M_⊙ yr⁻¹, in sharp contrast with the unity relation in more massive galaxies. We derive SFR calibrations for MIRI photometry, finding that a single, representative IR template improves the scatter. We additionally calibrate a UV+IR composite indicator, assuming energy balance, which does return a unity relation, with low scatter. Our examination of the mid-IR in our MS sample indicates that it is tracking the global obscuration fraction, making it a robust proxy for SFR down to our low mass end, log M_⋆/M_⊙∼9, and across the redshift range where MIRI probes the PAHs (0.3≲ z≲3). Finally, we examine the bright end not represented in SMILES, comparing the behavior of local and cosmic noon ultra luminous infrared galaxies to show that the robustness of using the IR as a SFR proxy extends from the faint to bright regimes.