OpenCodeReasoning: Advancing Data Distillation for Competitive Coding
Wasi Uddin Ahmad, Sean Narenthiran, Somshubra Majumdar, Aleksander Ficek, Siddhartha Jain, Jocelyn Huang, Vahid Noroozi, Boris Ginsburg
2025-04-04
Summary
This paper is about improving how AI learns to code by giving it a better training dataset.
What's the problem?
It's hard to train AI to code well because the training data is often not very good or not shared publicly.
What's the solution?
The researchers created a new, high-quality dataset for training AI to code and used it to achieve state-of-the-art results with smaller AI models. They prioritized the variety of problems over just having correct solutions.
Why it matters?
This work matters because it makes it easier for others to train AI models to code well and shows that having diverse training examples is important.
Abstract
Since the advent of reasoning-based large language models, many have found great success from distilling reasoning capabilities into student models. Such techniques have significantly bridged the gap between reasoning and standard LLMs on coding tasks. Despite this, much of the progress on distilling reasoning models remains locked behind proprietary datasets or lacks details on data curation, filtering and subsequent training. To address this, we construct a superior supervised fine-tuning (SFT) dataset that we use to achieve state-of-the-art coding capability results in models of various sizes. Our distilled models use only SFT to achieve 61.8% on LiveCodeBench and 24.6% on CodeContests, surpassing alternatives trained with reinforcement learning. We then perform analysis on the data sources used to construct our dataset, the impact of code execution filtering, and the importance of instruction/solution diversity. We observe that execution filtering negatively affected benchmark accuracy, leading us to prioritize instruction diversity over solution correctness. Finally, we also analyze the token efficiency and reasoning patterns utilized by these models. We will open-source these datasets and distilled models to the community.