The vast majority of tolerance research in political science takes place in the context of the US and other Western democracies (Marquart-Pyatt and Paxton 2007). This research defines tolerance as the willingness to grant rights characteristic of Western-style liberal democracies—such as freedoms of speech and association—to disliked groups. As a result, far less is known about tolerance in the context of autocracies and hybrid regimes, even though political theorists have shown that tolerance has evolved historically across sociopolitical contexts, albeit in differing forms (Murphy 1997). In this paper, we ask: What does tolerance mean in a non-Western context? Specifically, what kinds of behaviors are required to demonstrate it? And how robust is tolerance under authoritarianism? To answer these questions, we analyze data from an original, nationally-representative survey of 1000 Qatari citizens, conducted in February 2016. First, we probe the behaviors associated with tolerance through a question bank of ten possible actions, which range from avoidance of and politeness toward those who are different to allowing them to speak on television, teach one’s children, and occupy positions of power in society. These questions allow us both to move beyond the “rights”-dominant view peculiar to tolerance in the West, and to test hypotheses from political theory (Forst 2013) about minimal (such as “mere” non-interference) and maximal variations on tolerance (such as respect and recognition). The latter are increasingly seen by theorists as better suited to the needs of today’s more multicultural societies. Second, we also use a framing experiment to test the robustness of tolerance, drawing on the “slippage” hypothesis, which emphasizes the possible gap between abstract commitments to civil liberties and applications to concrete cases (Prothro and Grigg 1960). Half of the respondents begin with a question that asks them to identify groups they would not like to have as neighbors. Then, the tolerance question bank is presented, prefaced by asking respondents to focus on the concrete least-favored group(s) (thus meeting the “objection” criterion). To ascertain commitments to tolerance in the abstract, the other half receives the tolerance question bank first, so that respondents are not primed to think specifically about their least-favored groups. By combining the bank of questions on tolerant behaviors with a framing experiment, our research builds knowledge of the concept and robustness of tolerance in a contemporary authoritarian regime, and offers important lessons for state-led social engineering efforts to increase tolerance.