The zebrafish has been proposed for the analysis of the neurobiological and behavioral effects of alcohol in vertebrates. Significant behavioral changes induced by acute alcohol treatment, adaptation to chronic alcohol exposure, and withdrawal induced behavioral responses have all been shown in zebrafish. Previously, a flow-through Y-maze paradigm was proposed to directly measure alcohol preference or avoidance in zebrafish without the need to train learning-based place preference. Here, we first demonstrate that this Y-maze paradigm is capable of quantifying preference for a positive stimulus (the sight of conspecifics) and also the avoidance of a negative stimulus, a noxious olfactory cue, denatonium benzoate. Second, we test whether naïve zebrafish avoid alcohol upon first encountering this substance, and whether fish chronically exposed to alcohol show preference, or acutely alcohol treated fish show signs of intoxication leading to random choice. Our results demonstrate that acute alcohol treated fish exhibit enhanced immobility and perform at chance but chronic alcohol treated fish are not intoxicated and swim as well as naïve fish, a finding compatible with the known intoxicating effect of acute alcohol and the adaptation expected after chronic alcohol exposure. However, despite the general feasibility of the task, neither alcohol preference, nor alcohol avoidance could be detected in any of our treatment groups. We discuss the possible reasons why differential alcohol vs. freshwater choice was not found in this task and propose follow up experiments.