FullStack Bench: Evaluating LLMs as Full Stack Coders

TL;DR

This paper introduces FullStack Bench, a comprehensive dataset and evaluation framework for assessing large language models' capabilities across full-stack programming tasks in multiple languages, with a supporting sandbox tool.

Contribution

The paper presents a new multi-domain, multilingual code evaluation dataset and an execution sandbox, addressing limitations of existing benchmarks and enabling more realistic assessments.

Findings

FullStack Bench covers diverse application domains and 16 programming languages.

Experimental results show the effectiveness of our dataset and sandbox tool.

Our evaluation highlights the strengths and limitations of current LLMs in full-stack coding.

Abstract

As the capabilities of code large language models (LLMs) continue to expand, their applications across diverse code intelligence domains are rapidly increasing. However, most existing datasets only evaluate limited application domains. To address this gap, we have developed a comprehensive code evaluation dataset FullStack Bench focusing on full-stack programming, which encompasses a wide range of application domains (e.g., basic programming, data analysis, software engineering, mathematics, and machine learning). Besides, to assess multilingual programming capabilities, in FullStack Bench, we design real-world instructions and corresponding unit test cases from 16 widely-used programming languages to reflect real-world usage scenarios rather than simple translations. Moreover, we also release an effective code sandbox execution tool (i.e., SandboxFusion) supporting various programming languages and packages to evaluate the performance of our FullStack Bench efficiently. Comprehensive experimental results on our FullStack Bench demonstrate the necessity and effectiveness of our FullStack Bench and SandboxFusion.