universitas terbuka has been applying a test of final examination for measuring student learning outcomes. Type of final examination that developed is objective teset consisting of stem and option.option of test covering answer key and detractors. a good question should have detractors relatively homogenous, so that it is difficult to guessed by a student.criteria of good question can be determined by item anaysis with using classical test theory and modern test theory. unversitas terbuka has been doing item analysis base on classical test theory. This theory has the limitation characteristics of the index items were highly dependent on the group of test. unlike the case with modern test theory (item response theory). Item response theory is based on the ability of individual test. the purpose of this paper is to determine: 1 how the result of the final exam ite analysis using item response analysis, 2 how the comparison of the number of valid and invalid between item responseanalysis and classical teset analysis, and 3 how determine item characteristics of the items and estimate the ability of participants to the test item response using rasch model. the data used are student answers that took exams on subjects selected sample in 2009.1 and 2009.2. the total sample used by 8 subjects consisting of 16 sets of questions. the study results showed: 1 average difficulty level of the sample items was subject to items response items (rasch model) i sbetween o.618 to 1.417, 2 comparison of the results of the analysis of test items using classical analysis and the rasch model anaylysis is very significant, and 3 on the number of examinees who many (over 1.500) people, the number of percentage of items was valid by using items response analysis more than the classical analysis.