Factorio Learning Environment - Milestones: leaderboard
Factorio Learning Environment (FLE) milestones: how many discrete achievements an LLM agent reaches while writing Python programs to build a factory in Factorio, such as producing an item type for the first time or researching a technology. A count from the FLE leaderboard with no fixed maximum; higher is better.
Metric: Milestones reached. Source: jackhopkins.github.io. Saturation forecast: Estimated already saturated. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o | 30 | #357 |
| 2 | Claude 3.5 Sonnet | 30 | #362 |
| 3 | DeepSeek V3 | 22 | #329 |
| 4 | Gemini 2.0 Flash | 20 | #351 |
| 5 | Llama 3.3 70B | 16 | #578 |
| 6 | GPT-4o Mini | 14 | #598 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=factorio-learning-environment-milestones · How It Works · Data refreshed daily, snapshot 2026-11-05.