This manuscript introduces Argument Belief Models (AB models), a framework for reasoning about belief, knowledge, and evidence, integrating epistemic evidence logic and abstract argumentation. Building on the ideas of Topological Argumentation Models (TA models), that incorporate tools from argumentation theory into topological evidence logic, AB models distinguish between evidence (information states) and arguments (supporting justifications for evidence), thereby allowing for a more nuanced selection of rational beliefs. By defining an attack relation on arguments rather than directly on evidence, AB models clarify how relative argument strength influences belief formation. We formally define AB models, investigate their logical properties, and establish a connection between grounded belief in AB models and TA models. Finally, we show that under certain conditions, AB models align with TA models. This work contributes to a more refined logical understanding of belief formation and opens pathways for further research into structured epistemic reasoning based on argumentation.