Agora: Enhancing LLM Agent Reasoning Via Auction-Based Task Allocation
2026-07-10 • Artificial Intelligence
Artificial IntelligenceComputation and Language
AI summaryⓘ
The authors propose Agora, a new method to help large language models use different expert tools more effectively. Instead of just matching tasks roughly to tools, Agora runs auctions where tools bid to handle reasoning steps based on their true abilities, avoiding overconfidence. This approach helps assign the hardest problems to the best solvers and balances quality with cost. Tests showed Agora performs better than other methods when using the same set of tools.
large language modelsexpert modelstask allocationauction mechanismreasoning stepsperformance variabilitycost efficiencymodel routingcascade models
Authors
Kaiji Zhou, Ales Leonardis, Yue Feng
Abstract
Enhancing the reasoning capabilities of large language model (LLM) agents requires effective orchestration of diverse expert models and tools. However, existing frameworks typically call APIs based on coarse-grained matching between tasks and the functions of expert models or tools, while overlooking critical factors such as performance variability and cost efficiency among functionally similar alternatives. To address this, we propose Agora, a framework that introduces an incentive-compatible auction mechanism for dynamically allocating tasks to expert models and tools. By treating reasoning steps as tradeable items, Agora enables agents to bid based on their rectified competence-ensuring that critical logic is routed to the most capable solver rather than the most overconfident one. Evaluations across five benchmarks show that Agora improves over matched single-model, routing, and cascade baselines under comparable candidate pools, while exposing a controllable cost-quality trade-off through a single auction parameter.