SWE-Explore - Precision: leaderboard
SWE-Explore (2026) isolates repository exploration for coding agents: for each of 848 issues from 203 repositories in 10 languages the explorer returns its five most relevant code regions, scored against line-level context distilled from successful repair trajectories, here with each LLM driving the Mini-SWE-Agent scaffold. This board is line-level precision, the percent of returned lines inside that core context; higher is better.
Metric: Line-level precision (%; share of returned lines inside the core context) on 848 repository-exploration instances (issues from 203 repositories in 10 languages): given the repository and issue, the explorer returns five ranked code regions, scored against line-level core context distilled from independent successful repair trajectories; the Mini-SWE-Agent scaffold with each LLM; higher is better. Source: arxiv.org. Saturation forecast: Around September 2027. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.4 | 54.2 | #123 |
| 2 | Claude Sonnet 4.5 | 51.9 | #250 |
| 3 | GPT-5.4 Mini | 50.9 | #389 |
| 4 | Kimi K2.6 | 47.5 | #178 |
| 5 | Gemini 3 Pro | 42 | #137 |
| 6 | GLM-4.7 | 41.4 | #366 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=swe-explore-precision · How It Works · Data refreshed daily, snapshot 2026-11-13.