https://arxiv.org/abs/2505.02133 Enhancing LLM Code Generation: A Framework for Systematic Evaluation of Multi-Agent Collaboration and Runtime Debugging for Improved Accuracy, Reliability, and Latency. Published in IEEE Xplore as part of the 2025 IEEE 19th International Conference on AICT.