# Voyager: An Open-Ended Embodied Agent with Large Language Models

*Source* · `src-wang-2023-voyager` · status: **n/a**

| field | value |
|---|---|
| kind | Peer-reviewed paper |
| authors | Guanzhi Wang, Yuqi Xie, Yunfan Jiang, Ajay Mandlekar, Chaowei Xiao, Yuke Zhu, Linxi Fan, Anima Anandkumar |
| venue | Transactions on Machine Learning Research (2024) |
| date | 2023-05 |
| url | https://arxiv.org/abs/2305.16291 |
| arxiv | 2305.16291 |
| accessed | 2026-10-24 |

## Notes

Automatic curriculum, skill library of executable code, iterative prompting with environment feedback, execution errors, and self-verification; GPT-4. 3.3x unique items, 2.3x distance, tech-tree milestones up to 15.3x faster; 63 items in 160 iterations; GPT-4 15x cost of GPT-3.5; GPT-3.5 code gen 5.7x fewer items; hallucinated tasks (copper sword) and nonexistent API calls.

## Backlinks (computed)

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

## Citation

Guanzhi Wang, Yuqi Xie, Yunfan Jiang, Ajay Mandlekar, Chaowei Xiao, Yuke Zhu, Linxi Fan, Anima Anandkumar (2023-05). *Voyager: An Open-Ended Embodied Agent with Large Language Models*. Transactions on Machine Learning Research (2024) https://arxiv.org/abs/2305.16291

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