GraphLang Dataset — Commercial License
The Product
10,000,000 cross-language function pairs with aligned GraphLang IR output. The largest curated dataset of its kind. Used to train, fine-tune, and evaluate code intelligence models.
What You Get
benchmark_10m_results.json— 217,210,967 nodes normalizedbenchmark_1m_results.json— 21,721,250 nodesbenchmark_100k_results.json— 2,172,203 nodesbenchmark_dataset.json— 1,500 hand-curated pairs
All aligned: Python ↔ Java ↔ JavaScript ↔ GraphLang IR. 22.5x compression verified. 100% cross-language equivalence for same-logic functions.
Pricing
| License | Price | Usage |
|---|---|---|
| Academic | Free | Research, papers, non-commercial |
| Startup (<$1M revenue) | $5,000/year | Commercial use, AI training |
| Enterprise | $25,000/year | Unlimited commercial use, AI training, redistribution |
Buy Now
Contact: josu31.jas@gmail.com GitHub: https://github.com/cripto-bot/graphlang License: https://creativecommons.org/licenses/by-nc-sa/4.0/
To purchase a commercial license, email josu31.jas@gmail.com with subject "GraphLang Dataset License".
Payment via wire transfer or crypto.
"El dataset que demuestra que Python = Java = JavaScript al nivel semántico."