AIM: To determine the agreement and repeatability of the pupil measurement obtained with VIP-200 (Neuroptics), PowerRef II (Plusoptix), WAM-5500 (Grand Seiko) and study the effects of instrument design on pupillometry. METHODS: Forty patients were measured twice in low, mid and high mesopic. Repeatability was analyzed with the within-subject standard deviation (Sw) and paired t-tests. Agreement was studied with Bland-Altman plots and repeated measures ANOVA. Instrument design analysis consisted on measuring pupil size with PowerRef II simulating monocular and binocular conditions as well as with proximity cues and without proximity cues. RESULTS: The mean difference (standard deviation) between test-retest for low, mid and high mesopic conditions were, respectively: -0.09 (+/- 0.16), -0.05 (+/- 0.18) and -0.08 (+/- 0.23) mm for Neuroptics, -0.05 (+/- 0.17), -0.12 (+/- 0.23) and -0.17 (+/- 0.34) mm for WAM-5500, -0.04 (+/- 0.27), -0.13 (+/- 0.37) and -0.11 (+/- 0.28) mm for PowerRef II. Regarding agreement with Neuroptics, the mean difference for low, mid and high mesopic conditions were, respectively: -0.48 (+/- 0.35), -0.83 (+/- 0.52) and -0.38 (+/- 0.56) mm for WAM-5500, -0.28 (+/- 0.56),-0.70 (+/- 0.55) and -0.61 (+/- 0.54) mm for PowerRef II. The mean difference of binocular minus monocular pupil measurements was: -0.83 (+/- 0.87) mm; and with proximity cues minus without proximity cues was: -0.30 (+/- 0.77) mm. CONCLUSION: All the instruments show similar repeatability. In all illumination conditions, agreement of Neuroptics with WAM-5500 and PowerRef II is not good enough, which can be partially induced due to their open field design.