Selection of parents to maximize hybrid performance requires a clear understanding of the combining ability. 101 ICRISAT bred B-lines and two R-line testers were crossed using a line × tester design to develop 202 F1 hybrids and evaluated at four locations in an alpha-lattice design. The environments, hybrids and variance partitions due to lines, testers and their interaction (lines × testers) were found significant. Contribution of lines to the combining ability variance was very high for days to 50 % flowering (DF) (84.3 %), plant height (PH) (71.9 %), panicle girth (PG) (89.8 %), blast severity score (BS) (59.4 %), panicle yield (PY) (64 %), grain yield (GY) (61.8 %) and dry fodder yield (FY) (65.2 %). High percentage contribution of line × tester for panicle yield plot-1 (32.6), threshing percentage (43), grain yield (38.1), fodder yield (34.8) and blast score (28) revealed an interplay of additive and non-additive gene action in their inheritance, thereby emphasizing the potential for pedigree breeding in combination with harvesting the heterotic potential of crop for these traits. Several seed- parents were identified as good general combiners for DF (ICMB 88004, ICMB 04222), PH (ICMB 08888), PL (ICMB 09333, ICMB 09555) and GY (ICMB 93222, ICMB 94444, ICMB 07666). Inbreds ICMB 08666, ICMB 08888 and ICMB 09333 were identified good for biomass production. Hybrids like ICPH021, ICPH023, ICPH107, ICPH175, ICPH007 and ICPH083 were identified as top performing hybrids for different segments and agro-ecological zones. Grain and fodder yield showed a significant and positive correlation with GCA and SCA suggesting that combining ability estimates are reliable predictors of hybrid performance, aiding in the effective exploitation of heterosis.