TokenEconomizer: Experience-Driven Black-Box Cost Control for Coding Agents

TokenEconomizer, an experience-driven black-box cost-control system for coding agents, targets the token costs that frontier coding agents have sharply increased for AI-assisted software engineering. The work argues that token efficiency is typically evaluated on isolated tasks, while real-world developers use coding agents to continuously build, update, and maintain large-scale projects, a mismatch the system is designed to address.

Advances in frontier coding agents have sharply increased the token costs of AI-assisted software engineering. Their token efficiency is typically evaluated on isolated tasks, whereas real-world developers use them to continuously build, update, and maintain large-scale projects. This mismatch motiv