Increasing rice production on a sustainable basis is a major challenge to research for development in sub-Saharan Africa. However, apart from yield, little efforts have been made to benchmark related performance indicators (PIs) including net profit, labor productivity, and nitrogen and phosphorus use efficiencies (NUE and PUE). The objectives of this study are to (i) assess these PIs and their determinants, and (ii) identify areas needed to improve them. Data were collected from 2,060 farmers in irrigated lowland (IL), rainfed lowland (RL), and rainfed upland (RU) rice production systems across two agroecological zones (semi-arid and subhumid) in the 2018/19 wet season at 18 sites in Burkina Faso, Ghana, Nigeria, Senegal, and Tanzania. Average yields were 4,952, 2,988, and 1,816 kg ha −1 in IL, RL, and RU. Higher yield was associated with higher profit. Labor productivity was 55 kg day −1 in IL, 75 kg day −1 in RL, and 49 kg day −1 in RU. Labor productivity was generally low because of limited use of agricultural equipment. NUE and PUE differed across production systems: 17%, 42%, and 49% of farmers had below desirable NUE (<40 kg grain kg −1 N) in IL, RL, and RU. Meanwhile, 62%, 44%, and 34% of farmers had above desirable PUE (100–250 kg grain kg P −1 ) in IL, RL, and RU, indicating a need to increase P application to avoid soil P mining. Cluster analysis of PIs identified four groups across 24 site–production system combinations. The low-performing group had larger field size, lower values for five PIs, and poorer crop management practices. These results suggest opportunities for improving rice productivity in low-performing groups through introduction of integrated crop management options, including improved crop and nutrient management together with labor-saving technologies such as mechanization.