Over the past three decades, penicillin-resistant pneumococci have emerged worldwide. In addition, penicillin-resistant strains have also decreased susceptibility to other β-lactams (including cephalosporins) and these strains are often resistant to other antibiotic groups, making the treatment options much more difficult. Nevertheless, the present in vitro definitions of resistance to penicillin and cephalosporins in pneumococci could not be appropriated for all types of pneumococcal infections. Thus, current levels of resistance to penicillin and cephalosporin seem to have little, if any, clinical relevance in nonmeningeal infections (e.g., pneumonia or bacteremia). On the contrary, numerous clinical failures have been reported in patients with pneumococcal meningitis caused by strains with MICs ≥ 0.12 μg/ml, and penicillin should never be used in pneumococcal meningitis except when the strain is known to be fully susceptible to this drug. Today, therapy for pneumococcal meningitis should mainly be selected on the basis of susceptibility to cephalosporins, and most patients may currently be treated with high-dose cefotaxime (±) vancomycin, depending on the levels of resistance in the patient's geographic area. In this review, we present a practical approach, based on current levels of antibiotic resistance, for treating the most prevalent pneumococcal infections. However, it should be emphasized that the most appropriate antibiotic therapy for infections caused by resistant pneumococci remains controversial, and comparative, randomized studies are urgently needed to clarify the best antibiotic therapy for these infections