Oriel — AI & technology worth noticing, every day.

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

Paper · https://arxiv.org/abs/2610.22732

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 motivates us to evaluate agents over sequences of realistic coding tasks, where we find that agents exhibit recurring environment-dependent token costs that existing methods struggle to reduce. To minimize these costs, we introduce TokenEconomizer, an external controller that cuts environment-specific token costs over a sequence of tasks. These cuts come from determining (i) which operational micro-skills to inject, (ii) which persistently irrelevant tools to prune, and (iii) when to intervene in unproductive failure loops. Across Claude Code, Codex, and GitHub Copilot CLI paired with frontier models, TokenEconomizer reduces the average token cost by 18.4-30.9% with comparable task success. Savings hold in every repository under every tested agent-model setting and generalize to diverse domains such as sequential web development and security auditing, where costs fall by 6.4-21.2% depending on task structure. Our work establishes adaptive black-box cost control as a promising direction for reducing token costs of coding agents in realistic settings.

First seen by Oriel on .

AI summary

TokenEconomizer is an external controller that reduces coding agents’ token costs across sequences of realistic software tasks while maintaining comparable task success.

  • It controls costs by injecting operational micro-skills, pruning persistently irrelevant tools, and intervening in unproductive failure loops.
  • Across Claude Code, Codex, and GitHub Copilot CLI with frontier models, it cuts average token cost by 18.4–30.9%.
  • Savings hold across every tested repository and agent-model setting, and extend to sequential web development and security auditing with reductions of 6.4–21.2%.
  • The study evaluates agents over sequences of tasks to capture recurring environment-dependent costs missed by isolated-task evaluations.

#coding agents #token efficiency #agent control #software engineering

Generated by gpt-5.6-sol. Check the original for details.

Metric trends

Last 30 days · Up to 120 recorded samples per source and metric, with reposts kept separate · First and latest retained

Points

Score reported by the source.

  • Hugging Face Papers: 17, latest observation ; +14 over observed period since

Comments

Comment count reported by the source.

  • Hugging Face Papers: 2, latest observation ; 0 over observed period since

Aliases

SourceExternal IDURL
Hugging Face Papers2610.22732https://arxiv.org/abs/2610.22732 Read discussion