NR2B-selective inhibitors show lower side-effects in preclinical pain models than non-selective NMDA receptor (NMDAR) antagonists, but it is unclear whether the improved safety of NR2B-selective inhibitors is due to their subtype selectivity or to a unique mode of inhibition of the receptor. In this study, the analgesic effects of intracerebral bolus injections of conantokin peptides with different NMDAR subunit selectivity were determined in mice by the standard hot-plate test, and following stimuli with acetic acid, formalin and complete Freund's adjuvant (CFA). In the standard hot-plate model, con-G[S16Y], a NR2B-selective inhibitor, showed the highest analgesic activity among conantokin peptides tested. In the acetic acid- and CFA-induced pain models, con-G[S16Y] and, to a lesser extent, con-G exhibited higher analgesic activity compared to non-selective inhibitors, such as con-R[1-17]. In the formalin test, while all conantokin peptides could partially suppress the first phase response, only con-G[S16Y] and con-G significantly inhibited the second phase response and suppressed paw edema. Our results suggest that the antinociceptive action of the conantokins may be related to their NR2B-selectivity and that these peptides may be useful as both neurobiological tools for probing mechanisms of nociception and as therapeutic agents for pain relief.