Immune checkpoint therapy (ICT) causes durable tumor responses in a subgroup of patients. Profiling T cell receptor beta (TCRβ) repertoire structure in ICT responders and non-responders provides mechanistic insight into what constitutes an effective anti-tumor response, and could result in the development of predictive biomarkers of response to identify and stratify patients for ICT. To examine how the TCRβ repertoire dynamics contribute to ICT response, we utilized an established murine model that excludes variation in host genetics, environmental factors and tumor mutation burden, limiting variation between animals to naturally diverse TCRβ repertoires. Oligoclonal expansion of TCRβ clonotypes that corresponded with a low TCRβ diversity was observed in responding tumors prior to ICT. We modeled TCRβ cluster dynamics during ICT and found that select clonotypes expanded slower in responders compared to non- responders. Clonally expanded CD8+ tumor infiltrating T cells in non-responders exhibited a T cell exhaustion phenotype. We conclude that an early burst of clonal expansion followed by a contraction during ICT is associated with response.