The Stack v3 releases 114 TB of open code across 770 languages
anton_lozhkov · x · 2026-07-24
- The Stack v3 is out as the largest open source-code dataset released so far: 114 TB raw, 770 languages, 224M repositories, and about 5T deduplicated tokens.
- The training subset contains 15.9 TB of code across 713 languages from 173M repositories; the full corpus contains 113.7 TB across 770 languages from 224M repositories.
- Compared with v2, v3 reflects two more years of GitHub activity and includes inline file contents, making it fully self-contained for training.
- The release is positioned as a more open, reproducible, and transparent foundation for code model pretraining.
Related event: Hugging Face Releases The Stack v3: 114TB Largest Open Source Code Dataset(13 posts)→
More from Research
- Token Collapse: Study Finds Graph Soft Prompting Tokens in LLMs Fail to Encode Structure — Sapienza · 2026-10-27
- The 'sweet lesson' is real: two kinds of ML problems need opposite strategies — mike64_t · 2026-10-27
- PixGR: Pixel-space generative decoding for fast feed-forward 3D reconstruction — kwangmoo_yi · 2026-10-27
- UniAction: unified action conditioning lets one world model handle mixed action types — ZeYanjie · 2026-10-27
- Infini-AI Lab open-sources Hermit, a framework for agent harness evolution and continual learning — BeidiChen · 2026-10-27
- Meta FAIR's Self-Direction turns agent failures into reusable LoRA skills, +13.8pp on Terminal-Bench — karen_ullrich · 2026-10-27