Notwithstanding the introduction of education multilingual policies worldwide, testing and assessment procedures still rely almost exclusively on the monolingual construct. This paper describes a study, part of a larger project fostering a new multilingual education policy in Israeli schools, exploring bi/multilingual assessment. It included two types of second language learners--immigrants from the Former Soviet Union, who learn all their school subjects in Hebrew, and Arab students, whose school language of instruction is Arabic but who learn some subjects in Hebrew. The experimental groups received a bilingual version of a test (Hebrew-Russian; Hebrew Arabic) and the control groups a Hebrew-only version. In the Russian-speaking experimental group, students received significantly higher scores than in the control group, while no significant differences surfaced between the groups among the Arabic-speaking students. Yet, attitudes toward bilingual assessment, evaluated via questionnaires, think aloud protocols and focus groups, were highly positive in both groups, addressing the availability of L1 in the test as contributing greatly to a more relaxed and positive approach. Multilingual tools emerged as a fairer method of assessing knowledge for second-language learners, who cannot fully demonstrate their academic knowledge in L2, and their use is recommended as part of the new multilingual policy.