Benchmark
RepoQA
long_context
reasoning
code
text
多语言
RepoQA is a benchmark for evaluating long-context code understanding capabilities of Large Language Models through the Searching Needle Function (SNF) task, where LLMs must locate specific functions in code repositories using natural language descriptions. The benchmark contains 500 code search tasks spanning 50 repositories across 5 modern programming languages (Python, Java, TypeScript, C++, and Rust), tested on 26 general and code-specific LLMs to assess their ability to comprehend and navigate code repositories.
语言EN
满分1
参评模型2
模型排名
| 名次 | 模型 | 机构 | 分数 | 来源 |
|---|---|---|---|---|
| 1 | Phi-3.5-MoE-instruct | Microsoft | 85.0 | 来源 ↗ |
| 2 | Phi-3.5-mini-instruct | Microsoft | 77.0 | 来源 ↗ |