The Computer has occupied a comprehensive coverage, especially in education scopes, including in learning-teaching processes, testing, and evaluating. This research aimed to develop computerized adaptive testing (CAT) to measure physics higher-order thinking skills (HOTS), namely PhysTHOTS-CAT. The Research Development used the 4-D developmental model carrying the four phases of define, design, development, and dissemination (4D) developed by Thiagarajan. This testing instrument can give the item test based on the student's abilities. The research phases include (1) needs analysis and definition, (2) development design (3) development of CAT and assemble the test items into CAT, (4) validation by experts, and (5) feasibility try-out. The findings show that PhysTHOTS-CAT is valid to measure physics HOTS of the 10th-grade students of Senior High School according to 82.28% of teachers and students assessment on PhysTHOTS-CAT content and media. Therefore, it can conclude that PhysTHOTS-CAT can be used and feasible to measure physics HOTS of the 10th-grade students of the Senior High School.

Abstract-Assessment in learning is perceived to be effective when it can measure the abilitis of students. Assessment in the 21 st century emphasizes more on the realm of high-level abilitis. The objectives of this study were i) comparing Bloomian and Marizanoian higher order thinking skills (HOTS) in terms of construction and characteristics, and also ii) knowing the feasibility of applying the aforementioned two types of HOTS upon physics learning assessment. The Bloomian and Marzanoian HOTS instruments were developed by modifying the Wilson and Oriondo-Antonio models. The development models uutilized were 1) design test, 2) trial test, and 3) assembling test. The respondents for the trial were 737 public senior high school students in Yogyakarta Special Region. The results showed that 1) Bloomian HOTS consisting of 3 aspects, 8 sub-aspects, and 62 items with the most difficult aspect was create, while Marzanoian HOTS consisting of 5 aspects, 13 sub-aspects, and 63 items with the most difficult aspect was knowledge utilization. All items with subject material of Hooke's law, static fluids, temperature and heat, and optical devices were valid and fit with PCM, and also reliable; and 2) Bloomian and Marzanoian HOTS were feasibly applied in physics learning based on physics learning characteristics. Bloomian HOTS is better applied based on compliance with national curriculum demands, practicality, and efficiency.