CollabCoder: Plan-Code Co-Evolution via Collaborative Decision-Making for Efficient Code Generation

2026-04-15

AI Summary

Traditional multi-agent code generation approaches are limited by static planning, isolated execution, high computational overhead, and insufficient adaptability to complex tasks. This work proposes CollabCoder, a novel framework that introduces, for the first time, a collaborative decision-making mechanism between planning and coding modules. By dynamically scheduling the agent responsible for debugging phases, CollabCoder enables the co-evolution of plans and code. The framework supports multi-agent collaboration, adaptive debugging orchestration, and optimized invocation of large language model APIs. Evaluated on challenging benchmarks such as LiveCodeBench and xCodeEval, CollabCoder achieves performance gains of 11–20% while reducing API calls by an average of 4–10 per execution, substantially lowering computational costs.

Technology Category

• Multiagent Systems: Multiagent Planning
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language
• Planning, Routing, and Scheduling: Planning with Language Models

Application Category

• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Search and Retrieval-Augmented AI: Agentic search
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Automated code generation remains a persistent challenge in software engineering, as conventional multi-agent frameworks are often constrained by static planning, isolated execution, high computational overhead, and limited adaptability to complex tasks. This paper introduces CollabCoder, a novel Plan-Code Co-Evolution framework that improves code generation through dynamic multi-agent collaboration. The core idea is to design a collaborative decision-making process between the plan module and the code module to decide which module should be executed for the debugging process. Extensive experiments on widely used benchmarks demonstrate that CollabCoder consistently improves code quality and robustness across tasks. Importantly, CollabCoder achieves performance comparable to or exceeding current state-of-the-art methods while reducing computational overhead, with efficiency gains becoming more pronounced as benchmark difficulty increases. On the more challenging LiveCodeBench and xCodeEval benchmarks, our approach improves performance by 11-20% over strong baselines while reducing the number of API calls by an average of 4-10 per execution.

Problem

• automated code generation
• multi-agent frameworks
• computational overhead
• complex tasks
• code quality

Innovation

• Plan-Code Co-Evolution
• Collaborative Decision-Making
• Multi-Agent Code Generation
• Efficient Debugging
• Dynamic Planning