Cooperative Profiles Predict Multi-Agent LLM Team Performance in AI for Science Workflows

TL;DR

This paper demonstrates that cooperative behavior profiles derived from behavioral economics games can predict the performance of multi-agent LLM teams in scientific workflows, offering a diagnostic tool for assessing cooperation.

Contribution

It introduces a benchmarking approach linking game-based cooperative profiles of LLMs to their effectiveness in collaborative scientific tasks.

Findings

Game-derived cooperative profiles predict downstream scientific performance.

Models investing in team production outperform greedy strategies.

Cooperative disposition is a measurable property independent of general ability.

Abstract

Multi-agent systems built from teams of large language models (LLMs) are increasingly deployed for collaborative scientific reasoning and problem-solving. These systems require agents to coordinate under shared constraints, such as GPUs or credit balances, where cooperative behavior matters. Behavioral economics provides a rich toolkit of games that isolate distinct cooperation mechanisms, yet it remains unknown whether a model's behavior in these stylized settings predicts its performance in realistic collaborative tasks. Here, we benchmark 35 open-weight LLMs across six behavioral economics games and show that game-derived cooperative profiles robustly predict downstream performance in AI-for-Science tasks, where teams of LLM agents collaboratively analyze data, build models, and produce scientific reports under shared budget constraints. Models that effectively coordinate games and invest in multiplicative team production (rather than greedy strategies) produce better scientific reports across three outcomes, accuracy, quality, and completion. These associations hold after controlling for multiple factors, indicating that cooperative disposition is a distinct, measurable property of LLMs not reducible to general ability. Our behavioral games framework thus offers a fast and inexpensive diagnostic for screening cooperative fitness before costly multi-agent deployment.