Altering dietary carbohydrate (CHO) intake modulates fuel utilization during exercise. However, there has been no systematic evaluation of metabolic responses to graded changes in short-term (< 1 week) dietary CHO intake. Thirteen active men performed interval running exercise combined with isocaloric diets over 3 days before evaluation of metabolic responses to 60-min running at 65% V̇O2max on three occasions. Diets contained lower (LOW, 2.40 ± 0.66 g CHO.kg-1.d-1, 21.3 ± 0.5% of energy intake [EI]), moderate (MOD, 4.98 ± 1.31 g CHO.kg-1.d-1, 46.3 ± 0.7% EI), or higher (HIGH, 6.48 ± 1.56 g CHO.kg-1.d-1, 60.5 ± 1.6% EI) CHO. Pre-exercise muscle glycogen content was lower in LOW (54.3 ± 26.4 mmol.kg-1 wet weight [ww]) compared to MOD (82.6 ± 18.8 mmol.kg-1 ww) and HIGH (80.4 ± 26.0 mmol.kg-1 ww, P<0.001; MOD vs. HIGH, P=0.85). Whole-body substrate oxidation, systemic responses, and muscle substrate utilization during exercise indicated increased fat and decreased CHO metabolism in LOW (RER: 0.81 ± 0.01) compared to MOD (RER 0.86 ± 0.01, P = 0.0005) and HIGH (RER: 0.88 ± 0.01, P < 0.0001; MOD vs. HIGH, P=0.14). Higher basal muscle expression of genes encoding proteins implicated in fat utilization was observed in LOW. In conclusion, muscle glycogen availability and subsequent metabolic responses to exercise were resistant to increases in dietary CHO intake from ~5.0 to ~6.5 g CHO.kg-1.d-1 (46% to 61% EI), while muscle glycogen, gene expression and metabolic responses were sensitive to more marked reductions in CHO intake (~2.4 g CHO.kg-1.d-1, ~21% EI).