The active state conformation of a G-protein coupled receptor (GPCR) is influenced by the chemical structure and the efficacy of the bound ligand. Insight into the active state conformation as well as the activation pathway for ligands with different efficacies is critical in designing functionally specific drugs for GPCRs. Starting from the crystal structure of the beta 2-adrenergic receptor, we have used coarse grain computational methods to understand the modulation of the potential energy landscape of the receptor by two full agonists, two partial agonists, and an inverse agonist. Our coarse grain method involves a systematic conformational spanning of the receptor transmembrane helices followed by an energy minimization and ligand redocking in each sampled conformation. We have derived the activation pathways for several agonists and partial agonists, using a Monte Carlo algorithm, and these are in agreement with fluorescence spectroscopy measurements. The calculated pathways for the full agonists start with an energy downhill step leading to a stable intermediate followed by a barrier crossing leading to the active state. We find that the barrier crossing involves breaking of an interhelical hydrogen bond between helix5 and helix6, and polarization of the binding site residues by water facilitates the barrier crossing. The uphill step in the partial agonist salbutamol induced activation is distinct from full agonist norepinephrine, and originates from steric hindrance with the aromatic residues on helix6. Virtual ligand screening with the salbutamol-stabilized conformation shows enrichment of noncatechol agonists over the norepinephrine-stabilized conformation. Our computational method provides an unprecedented opportunity to derive hypotheses for experiments and also understand activation mechanisms in GPCRs.