rStar-Coder: Scaling Competitive Code Reasoning with a Large-Scale Verified Dataset

TL;DR

rStar-Coder introduces a large, verified dataset of 418K competitive programming problems and solutions, significantly enhancing large language models' code reasoning abilities through improved data quality and synthesis methods.

Contribution

The paper presents a novel dataset construction pipeline and a verified test case synthesis method to improve code reasoning in large language models.

Findings

rStar-Coder dataset leads to state-of-the-art performance on code reasoning benchmarks.

Models trained on rStar-Coder outperform previous models on LiveCodeBench and USACO.

The approach enables smaller models to achieve results comparable to larger models.

Abstract

Advancing code reasoning in large language models (LLMs) is fundamentally limited by the scarcity of high-difficulty datasets, especially those with verifiable input-output test cases necessary for rigorous solution validation at scale. We introduce rStar-Coder, which significantly improves LLM code reasoning capabilities by constructing a large-scale, verified dataset of 418K competition-level code problems, 580K long-reasoning solutions along with rich test cases of varying difficulty. This is achieved through three core contributions: (1) we curate competitive programming code problems and oracle solutions to synthesize new, solvable problems; (2) we introduce a reliable input-output test case synthesis pipeline that decouples the generation into a three-step input generation method and a mutual verification mechanism for effective output labeling; (3) we augment problems with high-quality, test-case-verified long-reasoning solutions. Extensive experiments on Qwen models (1.5B-14B) across various code reasoning benchmarks demonstrate the superiority of rStar-Coder dataset, achieving leading performance comparable to frontier reasoning LLMs with much smaller model sizes. On LiveCodeBench, rStar-Coder improves Qwen2.5-7B from 17.4% to an impressive 57.3%, and Qwen2.5-14B from 23.3% to 62.5%, surpassing o3-mini (low) by3.1%. On the more challenging USA Computing Olympiad, our 7B model achieves an average pass@1 accuracy of 16.15%, outperforming the frontier-level QWQ-32B. Code and the dataset will be released at https://github.com/microsoft/rStar.