Previous studies have assumed a linear relationship between inquiry-based teaching and student achievement in science. However this assumption may be questionable. Recent evidence on the effectiveness of inquiry-based teaching has yielded conflicting findings. To test the linearity assumption, the present study investigated the relation between inquiry-based teaching and achievement by taking into account the possible existence of nonlinear relations using Norwegian TIMSS 2015 data. A multilevel structural equation modeling analysis showed that the relationship was curvilinear. Inquiry-based teaching was positively correlated with achievement, but high frequency of inquiry activities was negatively related to achievement. Furthermore, we found that classroom SES did not affect the strength of the relationship between inquiry and achievement. These findings challenge the linearity assumption and contribute to explaining the conflicting evidence in earlier research as well as promote the effectiveness of inquiry-based teaching regardless of students’ socioeconomic backgrounds.