# Arctic-SnowCoder: Demystifying High-Quality Data in Code Pretraining

arXiv: https://arxiv.org/abs/2409.02326 · HTML: https://mlsanity.com/paper/2409.02326/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Builds on 15 of its 29 references that are in the map.
- First posted on arXiv 3 Sep 2024.

## Builds on (most cited first, 15 of 15)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://mlsanity.com/paper/1907.11692/) (2019)
- [Evaluating Large Language Models Trained on Code](https://mlsanity.com/paper/2107.03374/) (2021)
- [Program Synthesis with Large Language Models](https://mlsanity.com/paper/2108.07732/) (2021)
- [Enriching Word Vectors with Subword Information](https://mlsanity.com/paper/1607.04606/) (2016)
- [Is Your Code Generated by ChatGPT Really Correct? Rigorous Evaluation of Large Language Models for Code Generation](https://mlsanity.com/paper/2305.01210/) (2023)
- [DeepSeek-Coder: When the Large Language Model Meets Programming -- The Rise of Code Intelligence](https://mlsanity.com/paper/2401.14196/) (2024)
- [CodeGen: An Open Large Language Model for Code with Multi-Turn Program Synthesis](https://mlsanity.com/paper/2203.13474/) (2022)
- [BigCodeBench: Benchmarking Code Generation with Diverse Function Calls and Complex Instructions](https://mlsanity.com/paper/2406.15877/) (2024)
- [Magicoder: Empowering Code Generation with OSS-Instruct](https://mlsanity.com/paper/2312.02120/) (2023)
- [Language Models of Code are Few-Shot Commonsense Learners](https://mlsanity.com/paper/2210.07128/) (2022)
- [Granite Code Models: A Family of Open Foundation Models for Code Intelligence](https://mlsanity.com/paper/2405.04324/) (2024)
- [At Which Training Stage Does Code Data Help LLMs Reasoning?](https://mlsanity.com/paper/2309.16298/) (2023)
- [If LLM Is the Wizard, Then Code Is the Wand: A Survey on How Code Empowers Large Language Models to Serve as Intelligent Agents](https://mlsanity.com/paper/2401.00812/) (2024)
- [Top Leaderboard Ranking = Top Coding Proficiency, Always? EvoEval: Evolving Coding Benchmarks via LLM](https://mlsanity.com/paper/2403.19114/) (2024)

## Built on by (most cited first, 5 of 5)

- [SelfCodeAlign: Self-Alignment for Code Generation](https://mlsanity.com/paper/2410.24198/) (2024)
- [MegaMath: Pushing the Limits of Open Math Corpora](https://mlsanity.com/paper/2504.02807/) (2025)
- [CritiQ: Mining Data Quality Criteria from Human Preferences](https://mlsanity.com/paper/2502.19279/) (2025)
- [State-of-the-art Small Language Coder Model: Mify-Coder](https://mlsanity.com/paper/2512.23747/) (2025)
- [No Accidental Software Agent First Canonical Code for Human Code Entropy Reduction and 30 to 500 times Lower Frontier Model Requirements](https://mlsanity.com/paper/2606.14357/) (2026)

## Abstract (arXiv)

Recent studies have been increasingly demonstrating that high-quality data is crucial for effective pretraining of language models. However, the precise definition of "high-quality" remains underexplored. Focusing on the code domain, we introduce Arctic-SnowCoder-1.3B, a data-efficient base code model pretrained on 555B tokens through three phases of progressively refined data: (1) general pretraining with 500B standard-quality code tokens, preprocessed through basic filtering, deduplication, and decontamination, (2) continued pretraining with 50B high-quality tokens, selected from phase one by a BERT-style quality annotator trained to distinguish good code from random data, using positive examples drawn from high-quality code files, along with instruction data from Magicoder and StarCoder2-Instruct, and (3) enhanced pretraining with 5B synthetic data created by Llama-3.1-70B using phase two data as seeds, adapting the Magicoder approach for pretraining. Despite being trained on a limited dataset, Arctic-SnowCoder achieves state-of-the-art performance on BigCodeBench, a coding benchmark focusing on practical and challenging programming tasks, compared to similarly sized models trained on no more than 1T tokens, outperforming Phi-1.5-1.3B by 36%. Across all evaluated benchmarks, Arctic-SnowCoder-1.3B beats StarCoderBase-3B pretrained on 1T tokens. Additionally, it matches the performance of leading small base code models trained on trillions of tokens. For example, Arctic-SnowCoder-1.3B surpasses StarCoder2-3B, pretrained on over 3.3T tokens, on HumanEval+, a benchmark that evaluates function-level code generation, and remains competitive on BigCodeBench. Our evaluation presents a comprehensive analysis justifying various design choices for Arctic-SnowCoder. Most importantly, we find that the key to high-quality data is its alignment with the distribution of downstream applications.