Evaluating the Limitations of Local LLMs in Solving Complex Programming Challenges

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Matotek, Kadin, Cassel, Heather, Amiruzzaman, Md, Ngo, Linh B.
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866916957015506944
author Matotek, Kadin
Cassel, Heather
Amiruzzaman, Md
Ngo, Linh B.
author_facet Matotek, Kadin
Cassel, Heather
Amiruzzaman, Md
Ngo, Linh B.
contents This study examines the performance of today's open-source, locally hosted large-language models (LLMs) in handling complex competitive programming tasks with extended problem descriptions and contexts. Building on the original Framework for AI-driven Code Generation Evaluation (FACE), the authors retrofit the pipeline to work entirely offline through the Ollama runtime, collapsing FACE's sprawling per-problem directory tree into a handful of consolidated JSON files, and adding robust checkpointing so multi-day runs can resume after failures. The enhanced framework generates, submits, and records solutions for the full Kattis corpus of 3,589 problems across eight code-oriented models ranging from 6.7-9 billion parameters. The submission results show that the overall pass@1 accuracy is modest for the local models, with the best models performing at approximately half the acceptance rate of the proprietary models, Gemini 1.5 and ChatGPT-4. These findings expose a persistent gap between private, cost-controlled LLM deployments and state-of-the-art proprietary services, yet also highlight the rapid progress of open models and the practical benefits of an evaluation workflow that organizations can replicate on in-house hardware.
format Preprint
id arxiv_https___arxiv_org_abs_2509_15283
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Evaluating the Limitations of Local LLMs in Solving Complex Programming Challenges
Matotek, Kadin
Cassel, Heather
Amiruzzaman, Md
Ngo, Linh B.
Software Engineering
Artificial Intelligence
Machine Learning
Programming Languages
I.2.7; F.2.2; I.2.2
This study examines the performance of today's open-source, locally hosted large-language models (LLMs) in handling complex competitive programming tasks with extended problem descriptions and contexts. Building on the original Framework for AI-driven Code Generation Evaluation (FACE), the authors retrofit the pipeline to work entirely offline through the Ollama runtime, collapsing FACE's sprawling per-problem directory tree into a handful of consolidated JSON files, and adding robust checkpointing so multi-day runs can resume after failures. The enhanced framework generates, submits, and records solutions for the full Kattis corpus of 3,589 problems across eight code-oriented models ranging from 6.7-9 billion parameters. The submission results show that the overall pass@1 accuracy is modest for the local models, with the best models performing at approximately half the acceptance rate of the proprietary models, Gemini 1.5 and ChatGPT-4. These findings expose a persistent gap between private, cost-controlled LLM deployments and state-of-the-art proprietary services, yet also highlight the rapid progress of open models and the practical benefits of an evaluation workflow that organizations can replicate on in-house hardware.
title Evaluating the Limitations of Local LLMs in Solving Complex Programming Challenges
topic Software Engineering
Artificial Intelligence
Machine Learning
Programming Languages
I.2.7; F.2.2; I.2.2
url https://arxiv.org/abs/2509.15283