Intramuscular fat (IMF) % is an important measure of pork eating quality, with reduced IMF % linked to the selection of pigs with low backfat P2 thickness and more muscular genotypes over several decades. This experiment tested the association of IMF % from the M. longissimus thoracis et lumborum of 345 pigs at three abattoirs (Sites) with computed tomography (CT) determined carcass composition (% lean or fat), accredited abattoir measures of P2 backfat (mm) (Hennessy Grading Probe, AutoFom III and PorkScan Lite) and the lean % output from abattoir devices (Hennessy Grading Probe, AutoFom III and PorkScan Plus). There was a negative relationship between carcass CT lean % with IMF % at all Sites (P < 0.05), with the strength of this relationship varying between Sites. The strongest relationship was demonstrated at Site 1 (R2 0.30, RMSE 0.59), and across a 4 standard deviation range in the carcass lean % at each Site, resulted in a decrease in loin IMF % of 1.56, 0.44 and 0.80. There was a positive relationship of carcass CT fat % with IMF %, with similar, albeit slightly stronger relationship to IMF than with CT lean %. In contrast, only one of the three Sites with an accredited P2 measurement device demonstrated a significant association (P < 0.05) with IMF %. This is the first time the associations of IMF % with carcass CT composition has been assessed alongside that of P2 backfat measures from commercially deployed abattoir devices. Given P2 backfat is the key industry measure on which carcass value is determined in Australia, this experiment demonstrates the limitations that P2 has in measuring and monitoring the associations of carcass lean and fat % with IMF. Ideally, an independent measure of IMF % would allow for independent selection for lean % and IMF %, however there is currently no commercial device that can measure pork IMF %.