This study investigated the effect of carbohydrate (CHO) dose and composition on fuel selection during exercise, specifically exogenous and endogenous (liver and muscle) CHO oxidation. Ten trained males cycled in a double-blind randomized order on 5 occasions at 77% V˙O2max for 2 h, followed by a 30-min time-trial (TT) while ingesting either 60 g·h-1 (LG) or 75 g·h-113 C-glucose (HG), 90 g·h-1 (LGF) or 112.5 g·h-113 C-glucose-13 C-fructose ([2:1] HGF) or placebo. CHO doses met or exceed reported intestinal transporter saturation for glucose and fructose. Indirect calorimetry and stable mass isotope [13 C] tracer techniques were utilized to determine fuel use. TT performance was 93% "likely/probable" to be improved with LGF compared with the other CHO doses. Exogenous CHO oxidation was higher for LGF and HGF compared with LG and HG (ES > 1.34, P < 0.01), with the relative contribution of LGF (24.5 ± 5.3%) moderately higher than HGF (20.6 ± 6.2%, ES = 0.68). Increasing CHO dose beyond intestinal saturation increased absolute (29.2 ± 28.6 g·h-1 , ES = 1.28, P = 0.06) and relative muscle glycogen utilization (9.2 ± 6.9%, ES = 1.68, P = 0.014) for glucose-fructose ingestion. Absolute muscle glycogen oxidation between LG and HG was not significantly different, but was moderately higher for HG (ES = 0.60). Liver glycogen oxidation was not significantly different between conditions, but absolute and relative contributions were moderately attenuated for LGF (19.3 ± 9.4 g·h-1 , 6.8 ± 3.1%) compared with HGF (30.5 ± 17.7 g·h-1 , 10.1 ± 4.0%, ES = 0.79 & 0.98). Total fat oxidation was suppressed in HGF compared with all other CHO conditions (ES > 0.90, P = 0.024-0.17). In conclusion, there was no linear dose response for CHO ingestion, with 90 g·h-1 of glucose-fructose being optimal in terms of TT performance and fuel selection.