Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This study investigates the feasibility of using lower-cost models for repository exploration in coding-agent pipelines, aiming to maintain effective issue localization while reducing costs. By evaluating five explorer models on a diverse set of tasks, the authors find that cheaper models can achieve 78-94% of the localization quality of the highest-performing model while significantly decreasing agent time and token usage. The findings advocate for a modular approach to coding agents, allowing for tailored explorer selection based on specific downstream requirements.
Lower-cost models can achieve up to 94% of the localization quality of premium explorers while slashing agent time and token usage by up to 95%.
Repository exploration is a distinct and costly stage of coding-agent pipelines: before generating a patch, an agent must identify which repository files are likely to matter. We study whether this stage can be delegated to lower-cost models while retaining useful localization quality. Using our IssueLoc-Bench, we evaluate five explorer models under the same read-only interactive interface on 499 SWE-bench Verified-derived tasks and 500 tasks from 153 additional repositories. We measure early candidate discovery, top-three gold-file coverage, strict file-set recovery, agent time, and token usage, with paired instance-level uncertainty and repository-clustered sensitivity analysis. The highest-quality explorer leads across localization metrics, but substantially cheaper operating points emerge: depending on the model and evaluation arm, lower-cost explorers retain approximately 78-94% of the reference Hit@3 and 73-92% of its F1 while reducing mean agent time by 41-88% and token usage by 84-95%. The preferred operating point depends on how localization is consumed downstream: ranking and coverage metrics characterize recoverable candidate handoffs, whereas F1 and exact match characterize restrictive file gates. These results support treating repository exploration as an independently measurable and budgetable stage of modular coding agents, with explorer selection guided by the downstream handoff contract.