<![CDATA[The limited variety of items in the item bank remains an obstacle for Biology teachers in developing quality learning evaluations, even though the items used should be measured empirically for their quality and validity in order to provide comprehensive evaluation results. This study aims to examine the quality of senior high school Biology test items developed through Gemini AI. This study employed a quantitative descriptive approach by utilizing primary data from one class at SMA Pembangunan Laboratorium UNP. The research stages included item development using AI, content validation by experts, item tryout, answer collection and scoring, and item quality analysis. The results showed that the senior high school Biology evaluation items generated by Gemini AI and analyzed using the Rasch Model had fairly good quality, although improvements were still needed in several specific item quality indicators. These findings indicate that the use of Gemini AI can support teachers in developing and evaluating Biology test items more systematically. This study affirms that the use of Gemini AI has the potential to become an effective alternative in item quality analysis, while also providing practical implications for teachers to utilize similar technology in developing evaluation items in various subjects in order to obtain more reliable and higher-quality instruments. Keywords: Biology; Gemini AI; Item Quality; Rasch Model; Learning Evaluation]]>
Copyrights © 2026
