[12/14/2025] NOTE: We will no longer actively update this dataset. While this dataset is still functional and usable, we recommend you use the SWE-bench/SWE-smith-[lang] datasets.
Dataset Card
By SWE-bench, published under mit, revision ea6d7173829c.
[12/14/2025] NOTE: We will no longer actively update this dataset.
While this dataset is still functional and usable, we recommend you use the SWE-bench/SWE-smith-[lang] datasets.
For better maintainability and ease-of-use, we are maintaining language-specific datasets in lieu of this mono-repo.
The SWE-smith Dataset is a training dataset of 50137 task instances from 128 GitHub repositories, collected using the SWE-smith toolkit. It is the largest dataset to date for training software engineering agents.
All SWE-smith task instances come with an executable environment. To learn more about how to use this dataset to train Language Models for Software Engineering, please refer to the documentation.
Structure
default 59,136 rows
| Split | Rows | Size |
|---|---|---|
| train | 59,136 | 4.3 GB |
Details
- Repository
- SWE-bench/SWE-smith
- Publisher
- SWE-bench
- Task category
- Text generation
- Tags
- code, agents, software-engineering
- Size category
- 10K<n<100K
- Languages
- en
- Revision
- ea6d7173829c7ec8fa16c22055699ff2e9188091
- Last updated
- 2025-12-14
Files
13 files, 277.8 MB in total.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| data/train-00000-of-00011.parquet | Data | 7.9 MB | c9078e71de61 |
| data/train-00001-of-00011.parquet | Data | 12.4 MB | 3e8af6d9125b |
| data/train-00002-of-00011.parquet | Data | 4.1 MB | e17367bdd0fc |
| data/train-00003-of-00011.parquet | Data | 10.4 MB | dc8d53b2452e |
| data/train-00004-of-00011.parquet | Data | 116.0 MB | b6e9c84aaf14 |
| data/train-00005-of-00011.parquet | Data | 8.8 MB | e9e3025562ae |
| data/train-00006-of-00011.parquet | Data | 11.8 MB | 78dba1970cbd |
| data/train-00007-of-00011.parquet | Data | 16.2 MB | 27aa219993d8 |
| data/train-00008-of-00011.parquet | Data | 69.0 MB | e3cd42eae5ba |
| data/train-00009-of-00011.parquet | Data | 13.7 MB | 74dcce80661e |
| data/train-00010-of-00011.parquet | Data | 7.4 MB | 360c6b5b058d |
| README.md | Documentation | 1.8 KB | — |
| .gitattributes | Repository | 2.5 KB | — |
License and Download
- License
- mit
- Access
- No access gate
Released by SWE-bench through its official repository on Hugging Face. Read the license.