用合成数据训练编程模型，性能超越基线40分以上
X-Coder: Advancing Competitive Programming with Synthetic Tasks, Solutions, and Tests
- 全用合成任务、解法和测试用例完成微调到强化学习
- X-Coder-14B在LiveCodeBench上达67.5% avg@8
- 适合研究代码生成与合成数据训练的开发者
竞赛编程对代码大模型仍具挑战。尽管近期有所进展，但许多训练流程仍依赖稀缺的真实数据，引发可扩展性与近似重复基准污染的担忧。本文探索合成训练数据是否能支持竞赛编程的完整SFT至RL流程：不直接使用真实任务、解法或测试用例进行后训练。为此，我们生成了任务、经验证的解法及可靠的测试用例，作为强化学习的奖励信号。为提升奖励可靠性，引入双重验证策略，降低所选解法与测试输出中的噪声。基于高质量数据，训练了X-Coder模型系列。X-Coder-14B在LiveCodeBench v5上达到67.5% avg@8，v6上达63.4%，优于基线超40分。进一步分析揭示合成后训练的实用洞见，强调多样化任务、验证过的长思维链监督及以代码为中心的强化学习的价值。数据与模型已开源于https://github.com/JieWu02/X-Coder。
原文摘要 · Abstract (English)
Competitive programming remains challenging for code LLMs. Despite recent progress, many training pipelines still depend on scarce real-world data, raising concerns about scalability and near-duplicate benchmark contamination. In this paper, we examine whether synthetic training artifacts can support the complete SFT-to-RL cycle for competitive programming: no real-world tasks, solutions, or test cases are directly used for post-training. To this end, we synthesize tasks, verified solutions, and reliable test cases that serve as reward signals for reinforcement learning. To improve reward reliability, we introduce a dual-verification strategy that reduces noise in both selected solutions and test outputs. Using this high-quality data, we train the X-Coder model series. X-Coder-14B achieves 67.5\% avg@8 on LiveCodeBench v5 and 63.4\% on v6, outperforming its base model by over 40 points. Further analysis provides practical insights into synthetic post-training, highlighting the value of diverse tasks, verified long-CoT supervision, and code-centric reinforcement. Our data and models are released at https://github.com/JieWu02/X-Coder
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。