Saved in:
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2601.06953 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908801575157760 |
|---|---|
| author | Wu, Jie Li, Haoling Zhang, Xin Guo, Jiani Luo, Jane Liu, Steven Huang, Yangyu Chu, Ruihang Li, Scarlett Yang, Yujiu |
| author_facet | Wu, Jie Li, Haoling Zhang, Xin Guo, Jiani Luo, Jane Liu, Steven Huang, Yangyu Chu, Ruihang Li, Scarlett Yang, Yujiu |
| contents | Competitive programming poses a significant challenge for Code LLMs. While recent models have shown promise, they heavily rely on finite real-world data, raising concerns about scalability and contamination. In this paper, we investigate a critical question: Can we elevate models to expert-level reasoning performance using fully synthetic data? In response, we first observe that off-the-shelf synthesis methods yield suboptimal results in this domain. To address this, we systematically investigate the key factors governing synthetic data quality. Leveraging these findings, we significantly advance the feature-based synthesis paradigm via domain-specific evolution and a dual-verification strategy, promoting task solvability, solution correctness, and test accuracy. Using this high-quality synthetic data, we train the X-Coder model series under an SFT-then-RL paradigm. X-Coder-7B shows significant performance gains on the challenging LiveCodeBench v5 (62.9% avg@8) and v6 (55.8% avg@8), outperforming larger models trained on real-world data. Extensive analysis distills valuable insights into synthetic data scaling, the necessity of domain-adapted feature evolution, and code-centric reinforcement. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_06953 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | X-Coder: Advancing Competitive Programming with Fully Synthetic Tasks, Solutions, and Tests Wu, Jie Li, Haoling Zhang, Xin Guo, Jiani Luo, Jane Liu, Steven Huang, Yangyu Chu, Ruihang Li, Scarlett Yang, Yujiu Computation and Language Machine Learning Competitive programming poses a significant challenge for Code LLMs. While recent models have shown promise, they heavily rely on finite real-world data, raising concerns about scalability and contamination. In this paper, we investigate a critical question: Can we elevate models to expert-level reasoning performance using fully synthetic data? In response, we first observe that off-the-shelf synthesis methods yield suboptimal results in this domain. To address this, we systematically investigate the key factors governing synthetic data quality. Leveraging these findings, we significantly advance the feature-based synthesis paradigm via domain-specific evolution and a dual-verification strategy, promoting task solvability, solution correctness, and test accuracy. Using this high-quality synthetic data, we train the X-Coder model series under an SFT-then-RL paradigm. X-Coder-7B shows significant performance gains on the challenging LiveCodeBench v5 (62.9% avg@8) and v6 (55.8% avg@8), outperforming larger models trained on real-world data. Extensive analysis distills valuable insights into synthetic data scaling, the necessity of domain-adapted feature evolution, and code-centric reinforcement. |
| title | X-Coder: Advancing Competitive Programming with Fully Synthetic Tasks, Solutions, and Tests |
| topic | Computation and Language Machine Learning |
| url | https://arxiv.org/abs/2601.06953 |