Although the critical velocity (CV) protocol has been used to determine the aerobic capacity in rodents, there is a lack of studies that compare CV with maximal lactate steady state intensity (iMLSS) in mice. As a consequence, their physiological and molecular responses after exercise until exhaustion at CV intensity remain unclear. Thus, we aimed to compare and correlate CV with iMLSS in running mice, following different mathematical models for CV estimation. We also evaluated their physiological responses and muscle MCT1 and MCT4 after running until exhaustion at CV. Thirty C57BL/6J mice were divided into two groups (exercised-E and control-C). Group E was submitted to a CV protocol (4 days), using linear (lin1 and lin2) and hyperbolic (hyp) mathematical models to determine the distance, velocity, and time to exhaustion (tlim) of each predictive CV trial, followed by an MLSS protocol. After a running effort until exhaustion at CV intensity, the mice were immediately euthanized, while group C was euthanized at rest. No differences were observed between iMLSS (21.1 ± 1.1 m.min-1) and CV estimated by lin1 (21.0 ± 0.9 m.min-1, p = 0.415), lin2 (21.3 ± 0.9 m.min-1, p = 0.209), and hyp (20.6 ± 0.9 m.min-1, p = 0.914). According to the results, CV was significantly correlated with iMLSS. After running until exhaustion at CV (tlim = 28.4 ± 8,29 min), group E showed lower concentrations of hepatic and gluteal glycogen than group C, but no difference in the content of MCT1 (p = 0.933) and MCT4 (p = 0.123) in soleus muscle. Significant correlations were not found between MCT1 and MCT4 and tlim at CV intensity. Our results reinforce that CV is a valid and non-invasive protocol to estimate the maximal aerobic capacity in mice and that the content of MCT1 and MCT4 was not decisive in determining the tlim at CV, at least when measured immediately after the running effort.