@inproceedings{chen-etal-2026-paircoder,
    title = "{P}air{C}oder: Pair Programming-Inspired Two-Agent Collaboration for Code Generation",
    author = "Chen, Junhao  and
      Li, Xiang  and
      Xu, Yibin  and
      Cui, Yuehan  and
      Weng, Fangsheng  and
      Zhao, Hao  and
      Ma, Fei  and
      Tian, Qi",
    editor = "Liakata, Maria  and
      Moreira, Viviane P.  and
      Zhang, Jiajun  and
      Jurgens, David",
    booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {ACL} 2026",
    month = jul,
    year = "2026",
    address = "San Diego, California, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl/2026.findings-acl.149/",
    pages = "3043--3058",
    ISBN = "979-8-89176-395-1",
    abstract = "Large Language Models (LLMs) achieve strong results on code generation, but single model inference remains brittle on tasks that require iterative refinement. Existing multi agent frameworks improve reliability, yet they often incur substantial token and latency overhead. We introduce PairCoder, a framework that brings pair programming to autonomous LLM collaboration. PairCoder assigns one model to code generation and the other to review, and switches roles when repeated errors suggest that the current interaction has stalled. Across 13 LLMs on HumanEval, PairCoder consistently improves over single model inference. On eight representative backbones, it reaches 91.0{\%} pass@1 and improves over single model inference by up to 20.3{\%} while reducing token usage by 40{\%} to 70{\%} relative to multi agent baselines. Many heterogeneous pairings also outperform both constituent models, suggesting that the framework generalizes across model families. These results position PairCoder as an effective and deployment conscious alternative to heavier multi agent systems.Code is available at https://github.com/yisuanwang/PairCoder"
}