BACKGROUND: New computer numeric controlled (CNC)-milled frameworks for implant-supported prostheses have been introduced. However, no data are available on the precision of fit of these new frameworks. PURPOSE: The purpose of this study is to evaluate the precision of fit of a new CNC-milled framework technique (I-Bridge, Biomain AB, Helsingborg, Sweden) using Brånemark System (Nobel Biocare AB, Göteborg, Sweden) and NobelReplace (Nobel Biocare AB) system implants. MATERIALS AND METHODS: Ten test frameworks were fabricated for one master model for each implant system. Five additional frameworks were fabricated for five different models simulating clinical cases as controls (Brånemark System). The distortion of implant center point positions was measured in x-, y-, and z-axes and in three dimensions by using a contact-type coordinate measuring machine and a computer program developed specifically for this purpose. Mann-Whitney U-test was used to compare differences of distortion within and between the groups. RESULTS: The maximal distortion in arch width (x-axis) and curvature (y-axis) was within 71 and 55 microm for all frameworks, respectively. The mean distortion in absolute figures in x-, y-, z-axes and three dimensions was for "clinical control" frameworks 23, 26, 4, and 34 microm as compared with less than 12, 12, 2, and 17 microm for Brånemark and NobelReplace frameworks, respectively. Control frameworks showed significantly (p < .05) greater mean and range of distortions in x- and y-axes and in three dimensions compared with test frameworks. CONCLUSION: All measured frameworks presented signs of misfit, indicating that no framework had a "passive fit." Frameworks produced in a more routine clinical environment seem to present greater levels of distortion as compared with frameworks produced in a strict test situation. However, all measured frameworks presented levels of precision of fit within limits considered to be clinically acceptable in earlier studies of frameworks placed on abutments.