Current tuberculosis treatment is unsatisfactory due to its length (six months), its complexity (four drugs), and its ineffectiveness against multidrug-resistant strains. However, after more than 40 years without any new compounds, numerous new anti-tuberculosis drugs are currently in development. Some are derivatives of already known families, such as fluoroquinolones and rifamycins, while others belong to new families, such as diarylquinolines. These molecules offer the prospect of simplifying treatment for tuberculosis caused by susceptible bacilli (shorter total duration, weekly administration of medications). Furthermore, the new classes of anti-tuberculosis drugs may also be effective against multidrug- and extensively drug-resistant bacilli. (R.A.)