Scaling Small Agents Through Strategy Auctions

2026-02-02

AI Summary

This work addresses the limited scalability of small language models on complex tasks and the challenge of balancing cost and performance in existing multi-agent routing mechanisms. The authors propose SALE, a novel strategy auction framework that introduces a freelance-market-inspired coordination paradigm among agents. In SALE, agents bid with executable strategy proposals, and dynamic task allocation is achieved through a cost-value scoring mechanism coupled with shared auction memory. Notably, the framework enables test-time self-optimization without requiring additional training of a dedicated router. Evaluated on deep reasoning and code generation tasks, SALE reduces reliance on the largest model by 53%, lowers total inference cost by 35%, and surpasses the pass@1 performance of the largest model with only marginal computational overhead.

Technology Category

• Planning, Routing, and Scheduling: Planning with Language Models
• Multiagent Systems: Mechanism Design
• Search and Optimization: Learning to Search

Application Category

• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Search and Retrieval-Augmented AI: Agentic search
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Small language models are increasingly viewed as a promising, cost-effective approach to agentic AI, with proponents claiming they are sufficiently capable for agentic workflows. However, while smaller agents can closely match larger ones on simple tasks, it remains unclear how their performance scales with task complexity, when large models become necessary, and how to better leverage small agents for long-horizon workloads. In this work, we empirically show that small agents'performance fails to scale with task complexity on deep search and coding tasks, and we introduce Strategy Auctions for Workload Efficiency (SALE), an agent framework inspired by freelancer marketplaces. In SALE, agents bid with short strategic plans, which are scored by a systematic cost-value mechanism and refined via a shared auction memory, enabling per-task routing and continual self-improvement without training a separate router or running all models to completion. Across deep search and coding tasks of varying complexity, SALE reduces reliance on the largest agent by 53%, lowers overall cost by 35%, and consistently improves upon the largest agent's pass@1 with only a negligible overhead beyond executing the final trace. In contrast, established routers that rely on task descriptions either underperform the largest agent or fail to reduce cost -- often both -- underscoring their poor fit for agentic workflows. These results suggest that while small agents may be insufficient for complex workloads, they can be effectively"scaled up"through coordinated task allocation and test-time self-improvement. More broadly, they motivate a systems-level view of agentic AI in which performance gains come less from ever-larger individual models and more from market-inspired coordination mechanisms that organize heterogeneous agents into efficient, adaptive ecosystems.

Problem

• small language models
• task complexity
• agentic AI
• workload scaling
• agent coordination

Innovation

• Strategy Auctions
• Small Language Models
• Agentic AI
• Task Routing
• Cost-Efficient Coordination