EET: Experience-Driven Early Termination for Cost-Efficient Software Engineering Agents

TL;DR

EET tackles the high cost of SE agents by leveraging past experience to guide early termination during patch generation and selection. It builds a compact experience base from historical resolutions and retrieves relevant signals via TF-IDF to decide when to stop iterations. Empirical results on SWE-bench Verified across multiple agents and backends show substantial cost reductions (31.8% avg) with minimal impact on task success, and the approach is general enough to apply across diverse SE agents. The work offers a practical method to make automated SE agents more cost-efficient while preserving performance.

Abstract

Software engineering (SE) agents powered by large language models are increasingly adopted in practice, yet they often incur substantial monetary cost. We introduce EET, an experience-driven early termination approach that reduces the cost of SE agents while preserving task performance. EET extracts structured experience from prior issue-resolution executions and leverages it to guide early termination during patch generation and selection, reducing unproductive iterations. We evaluate EET on the SWE-bench Verified benchmark across three representative SE agents. EET consistently reduces total cost by 19%-55% (32% on average), with negligible loss in resolution rate (at most 0.2%). These efficiency gains are achieved, on average, by identifying early-termination opportunities for 11% of issues and reducing API calls, input tokens, and output tokens by 21%, 30%, and 25%, respectively. We release the code, prompts, and data at https://github.com/EffiSEAgent/EET.

Figures (2)

• Figure 1: Overview of EET.
• Figure 2: Experience object of the django__django-16910 issue.