Mycobacterium tuberculosis (Mtb) causes human tuberculosis, and more people die of it than of any other pathogen in the world. Immunodominant antigens elicit the large majority of T cells during an infection, making them logical vaccine candidates. Yet, it is still unknown whether these immunodominant antigen-specific T cells recognize Mtb-infected cells. Two immunodominant antigens, TB10.4 and Ag85b, have been incorporated into vaccine strategies. Surprisingly, mice vaccinated with TB10.4 generate TB10.4-specific memory CD8+ T cells but do not lead to additional protection compared to unvaccinated mice during TB. Ag85b-specific CD4+ T cells are also generated during vaccination, but the literature on whether these cells recognize Mtb-infected cells is also inconsistent. We demonstrate that TB10.4-specific CD8+ T cells do not recognize Mtb-infected cells. However, under the same conditions, Ag85b-specific CD4+ T cells recognize Mtb-infected macrophages and inhibit bacterial growth. In contrast, polyclonal CD4+ and CD8+ T cells from the lungs of infected mice can specifically recognize Mtb-infected macrophages, suggesting macrophages present antigens other than the immunodominant TB10.4. The antigen location may also be critical for presentation to CD8+ T cells, and live Mtb may inhibit antigen presentation of TB10.4. Finally, we propose that TB10.4 is a decoy antigen as it elicits a robust CD8+ T cell response that poorly recognizes Mtb-infected macrophages, allowing Mtb to evade host immunity.