Agonists at mu opioid receptors relieve acute pain, however, their long-term use is limited by side effects, which may involve beta-arrestin2. Agonists biased against beta-arrestin2 recruitment may be advantageous. However, the classification of bias may be compromised by assays utilising overexpressed mu receptors which overestimate efficacy for G-protein activation. There is a need for re-evaluation with restricted receptor availability to determine accurate agonist efficacies. We depleted mu receptor availability in PathHunter CHO cells using the irreversible antagonist, beta-funaltrexamine (beta-FNA), and compared efficacies and apparent potencies of twelve agonists, including several previously reported as biased, in beta-arrestin2 recruitment and cAMP assays. With full receptor availability all agonists had partial efficacy for stimulating beta-arrestin2 recruitment relative to DAMGO, while only TRV130 and buprenorphine were partial agonists as inhibitors of cAMP accumulation. Limiting receptor availability by prior exposure to beta-FNA (100 nM) revealed morphine, oxycodone, PZM21, herkinorin, U47700, tianeptine and U47931e are also partial agonists in the cAMP assay. The efficacies of all agonists, except SR-17018, correlated between beta-arrestin2 recruitment and cAMP assays, with depleted receptor availability in the latter. Furthermore, naloxone and cyprodime exhibited non-competitive antagonism of SR-17018 in the beta-arrestin2 recruitment assay. Limited antagonism by naloxone was also non-competitive in the cAMP assay, while cyprodime was competitive. Furthermore, SR-17018 only negligibly diminished beta-arrestin2 recruitment stimulated by DAMGO (1 mu M), whereas fentanyl, morphine and TRV130 all exhibited the anticipated competitive inhibition. The data suggest that SR-17018 achieves bias against beta-arrestin2 recruitment through interactions with mu receptors outside the orthosteric agonist site. This article is part of the Special Issue on "Ligand Bias".