SidConArena: An Environment Evaluating Agents in Open-Ended,Positive-Sum Bargaining Game

2026-06-24

AI Summary

This work addresses the limitations of existing evaluation environments in assessing large language models’ capabilities in negotiation, resource allocation, and long-term planning within open-ended, positive-sum settings. To this end, the authors introduce a multi-agent economic simulation framework that integrates three sequential phases: natural language negotiation, deterministic production transformation, and sealed-bid auctions, enabling open-ended positive-sum interactions under mixed motivations. The framework innovatively combines structured rules with free-form linguistic interaction through a phase-aware agent scheduling mechanism and a neuro-symbolic action interface. It is formalized as a finite-horizon partially observable stochastic game, enhanced with asynchronous execution and a structured observation space. Experiments demonstrate that state-of-the-art large models perform better in both homogeneous and heterogeneous configurations, yet still exhibit pervasive issues such as resource misjudgment, passive bargaining behavior, and insufficient long-term investment planning.

Abstract

Evaluating LLM agents requires dynamic environments that go beyond static reasoning and zero-sum games. Real-world economic interaction is often open-ended and mixed-motive: agents must negotiate, create positive-sum surplus, compete for scarce assets, and plan under delayed returns. We introduce SidConArena, a new benchmark framework for evaluating LLM agents in open-ended, positive-sum bargaining. SidConArena formalizes a multi-player economy as a finite-horizon partially observable stochastic game with three coupled phases: natural-language negotiation with binding trades, deterministic converter-based production, and sealed-bid auctions for long-term assets. The framework combines structured observations, phase-aware agent dispatching, a neural-symbolic action interface, and asynchronous execution, enabling free-form interaction while preserving rule-grounded evaluation. Across homogeneous and heterogeneous tournaments, stronger frontier models achieve higher economic outcomes, yet agents still misvalue resources, bargain passively, and remain limited in long-horizon investment planning.

Problem

• open-ended bargaining
• positive-sum game
• LLM agent evaluation
• economic interaction
• multi-player negotiation

Innovation

• positive-sum bargaining
• partially observable stochastic game
• neural-symbolic action interface
• asynchronous execution
• multi-agent economy