Saved in:
Bibliographic Details
Main Authors: Hopkins, Jack, Bakler, Mart, Khan, Akbir
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2503.09617
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929757020487680
author Hopkins, Jack
Bakler, Mart
Khan, Akbir
author_facet Hopkins, Jack
Bakler, Mart
Khan, Akbir
contents Large Language Models (LLMs) are rapidly saturating existing benchmarks, necessitating new open-ended evaluations. We introduce the Factorio Learning Environment (FLE), based on the game of Factorio, that tests agents in long-term planning, program synthesis, and resource optimization. FLE provides exponentially scaling challenges -- from basic automation to complex factories processing millions of resource units per second. We provide two settings: (1) lab-play consisting of eight structured tasks with fixed resources, and (2) open-play with the unbounded task of building the largest factory on an procedurally generated map. We demonstrate across both settings that models still lack strong spatial reasoning. In lab-play, we find that LLMs exhibit promising short-horizon skills, yet are unable to operate effectively in constrained environments, reflecting limitations in error analysis. In open-play, while LLMs discover automation strategies that improve growth (e.g electric-powered drilling), they fail to achieve complex automation (e.g electronic-circuit manufacturing).
format Preprint
id arxiv_https___arxiv_org_abs_2503_09617
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Factorio Learning Environment
Hopkins, Jack
Bakler, Mart
Khan, Akbir
Multiagent Systems
Computation and Language
Machine Learning
Large Language Models (LLMs) are rapidly saturating existing benchmarks, necessitating new open-ended evaluations. We introduce the Factorio Learning Environment (FLE), based on the game of Factorio, that tests agents in long-term planning, program synthesis, and resource optimization. FLE provides exponentially scaling challenges -- from basic automation to complex factories processing millions of resource units per second. We provide two settings: (1) lab-play consisting of eight structured tasks with fixed resources, and (2) open-play with the unbounded task of building the largest factory on an procedurally generated map. We demonstrate across both settings that models still lack strong spatial reasoning. In lab-play, we find that LLMs exhibit promising short-horizon skills, yet are unable to operate effectively in constrained environments, reflecting limitations in error analysis. In open-play, while LLMs discover automation strategies that improve growth (e.g electric-powered drilling), they fail to achieve complex automation (e.g electronic-circuit manufacturing).
title Factorio Learning Environment
topic Multiagent Systems
Computation and Language
Machine Learning
url https://arxiv.org/abs/2503.09617