The introduction of highly active antiretroviral therapy (HAART) combining potent drugs that can inhibit reverse transcriptase, integrase and protease activities has changed the natural history of the human immunodeficiency virus (HIV) type 1 disease. Unfortunately, poor penetrability into different anatomic compartments, toxicity and drug resistance are some of the problems related to their prolonged use. The ability of HIV to mutate and become resistant, along with the ongoing viral replication during HAART, can lead to the emergence of independently evolving viral strains in different anatomic compartments (i.e., brain, testes, lymph nodes, etc.). In addition, HAART predominantly effects the viral replication in the activated or differentiating CD(+) T lymphocytes, but appears to have a very limited effect on HIV-1 preintegration complexes in the latently infected cells. Existing drug therapies do not eliminate these viral reservoirs, nor do they prevent their formation. New strategies are needed for eliminating protected areas of HIV-1 in vivo. Therefore, the persistence of latent HIV-1 reservoirs is the principal barrier in the complete eradication of HIV-1 infection in patients by antiretroviral therapy at present. African non-human primates (NHPs) naturally infected with various simian immunodeficiency viruses (SIVs) appear not to develop immunodeficiency or AIDS, whereas Asian NHPs, which are unnatural hosts, infected with SIVs, as well humans infected with HIV-1, will nearly always develop progressive loss of CD(+) T lymphocytes and a gradual destruction of immune functions. Understanding the difference in the host responses between natural and unnatural hosts, and deciphering which host factors are responsible for the non-pathogenic course of natural SIV infections, would be valuable in developing more-effective treatment or prevention strategies for HIV/AIDS. A number of factors encoded by host cells have been identified that appear to play critical roles in the SIV infection process. Two of these factors, TRIM5alpha (a member of a large family of proteins known as the TRIM proteins) and cellular apolipoprotein B mRNA-editing enzyme-catalytic polypeptide-like-3G (APOBEC3G) have been recently identified. APOBEC3G genes belong to a family of primate genes that produce enzymes (in this case, APOBEC3G) that 'edit' RNA by replacing cytosine with guanine into viral particles as the virus undergoes reverse transcription in the cytoplasm of the host cell. HIV-1, in turn, counters with a protein called viral infectivity factor (Vif), which binds to the APOBEC3G enzyme that degrades it. Several other blocking factors have been described, including lentiviral blocking factor (Lv)1 and 2. These factors appear to block the infection at a postentry step; after reverse transcription has occurred, but before proviral integration. Thus, it is crucial to understand the molecular mechanisms involved in the establishment, maintenance and reactivation of lentiviral latency. This review presents various models of HIV-1 latency and forward a new unified model of lentiviral latency.