Abstract
As large language models (LLMs) are increasingly used in decision support, it is important to understand whether their choices under uncertainty exhibit stable and interpretable behavioral regularities. Human decision-making combines relatively persistent risk preferences with context-dependent adjustment, yet it remains unclear whether analogous behavioral structure can be observed in LLM-based decision systems.
Here we examine this question using a controlled multi-model framework based on no-limit Texas Hold'em, where behavior is quantified by Participation, measuring voluntary engagement in uncertain opportunities, and Proactiveness, measuring pre-flop risk escalation. Across homogeneous self-play and heterogeneous mixed-model interactions, frontier LLMs exhibit stable, model-specific risk profiles, forming a spectrum from conservative to aggressive decision styles. These profiles remain largely robust under changing opponent composition, while the most conservative and most aggressive models diverge further in mixed settings.
Under global risk pressure and personal resource constraint, models adapt in structured but heterogeneous ways, ranging from broad behavioral contraction to selective de-escalation and near-invariant behavior. These findings suggest that LLMs differ not only in baseline risk disposition, but also in the risk signals they respond to and the flexibility with which they adjust, providing a behavioral basis for auditing risk-sensitive decision-making in interactive settings.
Our code is publicly available at: GitHub - AgentTexasPoker.
Blogger's Review: This paper showcases the potential of large language models in decision support, particularly their risk assessment capabilities in complex environments. By employing a poker framework, the researchers reveal the adaptive decision-making behaviors of LLMs under various risk scenarios, advancing our understanding of model behaviors. This study lays a significant theoretical and practical foundation for future applications in dynamic decision-making.