StarCoderData
BigCode ProjectBigCode's filtered code corpus of roughly 250 billion tokens drawn from The Stack v1, covering source files alongside issues and notebooks, used to train StarCoder. It became a standard open code-pretraining component, documented in TinyLlama, IBM Granite Code, ALIA, Apertus, Marin and OLMoE.
Derived from The Stack v1; the map's the-stack-v2 record covers the family's current version.
Openness
3 high confidence- license
- permissive(SPDX per file)
- gated
- terms-acceptance
- dataset_card
- present
Terms-gated like its parent The Stack v1; permissively licensed source files.
- https://arxiv.org/abs/2305.06161 recorded 2026-09-29
The StarCoder paper abstract says "StarCoderBase is trained on 1 trillion tokens sourced from The Stack, a large collection of permissively licensed GitHub repositories" (the parent corpus of StarCoderData).
- https://huggingface.co/datasets/bigcode/starcoderdata recorded 2026-09-29
The dataset card page is gated with "You need to agree to share your contact information to access this dataset", describes the corpus as "This is the dataset used for training StarCoder and StarCoderBase", and quotes The Stack terms that "The Stack is a collection of source code from repositories with various licenses".
Adoption
3 high confidenceAdoption is measured by Hugging Face downloads of bigcode/starcoderdata, the only usage channel on record for StarCoderData.
- https://huggingface.co/api/datasets/bigcode/starcoderdata recorded 2026-08-13
28,873 downloads in the trailing 30 days for bigcode/starcoderdata
Capability
3 high confidenceStarCoderData is a proven pretraining component used in TinyLlama, Granite Code, ALIA and OLMoE, where it supplies 101 billion tokens to the OLMoE mix. It is drawn from The Stack v1, though, and has been functionally superseded by The Stack v2, which is four times its size, and by Stack-Edu.
- https://arxiv.org/abs/2402.19173 recorded 2026-09-29
The StarCoder2 paper abstract says the dataset "results in a training set that is 4x larger than the first StarCoder dataset", so the Stack v2 lineage is larger than StarCoderData.
- https://huggingface.co/datasets/allenai/OLMoE-mix-0924 recorded 2026-08-13
Mix statistics - Starcoder contributes 101B tokens (78.7M docs) to the OLMoE-mix-0924 pretraining mix
- https://huggingface.co/datasets/allenai/OLMoE-mix-0924/raw/main/README.md recorded 2026-09-29
The OLMoE-mix card lists Starcoder as a component with "| [Starcoder](https://huggingface.co/datasets/bigcode/starcoderdata) | 101 B" tokens in the pretraining mix.
Verified 2026-09-27