# Factorio Learning Environment

*Source* · `src-hopkins-2025-factorio-learning-environment` · status: **n/a**

| field | value |
|---|---|
| kind | Preprint |
| authors | Jack Hopkins, Mart Bakler, Akbir Khan |
| venue | arXiv |
| date | 2025-03-06 |
| url | https://arxiv.org/abs/2503.09617 |
| arxiv | 2503.09617 |
| accessed | 2026-10-24 |

## Notes

Lab-play (eight structured tasks) and open-play (largest factory on a procedural map). LLM agents found strategies such as electric drilling but failed at complex automation such as electronic circuits; weak spatial reasoning and error recovery.

## Backlinks (computed)

**cited by:** [LLM-written skill libraries (Voyager)](../technique/voyager-skill-libraries.md)

## Citation

Jack Hopkins, Mart Bakler, Akbir Khan (2025-03-06). *Factorio Learning Environment*. arXiv https://arxiv.org/abs/2503.09617

---
Provenance: created in run `2026-10-24-techniques-llm-agents`.