Artificial Analysis at least reports the results with fallback to an inferior model. So presumably Opus 5, and the score should be between Mythos 5.1 and that other model.