---
type: "Research"
title: "ReAct (2022)"
description: "Interleaving reasoning traces with actions improves task success."
resource: "https://arxiv.org/abs/2210.03629"
tags: [agents, tool-use, research]
generated: { by: human:crpage, at: 2026-07-09T09:44:00Z }
status: stable
sources: [{ id: primary, resource: "https://arxiv.org/abs/2210.03629" }]
---

Showed that **interleaving reasoning with actions** improves interpretability and long-horizon task performance (QA, ALFWorld, WebShop). Shifted the question from *can a model answer?* to *can a system reason and act?*