Abstract. The cube construction task represents a novel format in the assessment of spatial ability through mental cube rotation tasks. Instead of selecting the correct answer from several response options, respondents construct their own response in a computerized test environment, leading to a higher demand for spatial ability. In the present study with a sample of 146 German high-school students, we tested an approach to manipulate the item difficulties in order to create items with a greater difficulty range. Furthermore, we compared the cube task in a distractor-free and a distractor-based version while the item stems were held identical. The average item difficulty of the distractor-free format was significantly higher than in the distractor-based format ( M = 0.27 vs. M = 0.46) and the distractor-free format showed a broader range of item difficulties (.02 ≤ pi ≤ .95 vs. .37 ≤ pi ≤ .63). The analyses of the test results also showed that the distractor-free format had a significantly higher correlation with a broad intelligence test ( r = .57 vs. r = .17). Reasons for the higher convergent validity of the distractor-free format (prevention of response elimination strategies and the broader range of item difficulties) and further research possibilities are discussed.