ModelRefs / ReAct — Prompt Pattern

ReAct — Prompt Pattern

Interleave Reasoning and Acting (tool calls) in a single loop. Model alternates Thought → Action → Observation until it decides to answer.

Overview

Model alternates Thought → Action → Observation until it decides to answer. Foundational tool-use loop.

When to use it: You need an agent that can use tools and reason about their outputs.

Pattern details

Pattern classagent
Difficultyintermediate
Also known asreact agent, reasoning + acting
Last reviewed2026-06-07

Known failure modes

  • Runaway loops — Agent keeps calling tools forever. Mitigation: Set max steps; require Final Answer trigger.
  • Format drift — Agent stops emitting Action/Observation tokens. Mitigation: Validate format each step; retry on drift.

When not to use it

  • Using ReAct when a single function call would do.

Continue your research

Use these connected ModelRefs sections to compare alternatives, inspect implementation paths, and review the evidence and governance boundaries relevant to ReAct — Prompt Pattern.

Frequently asked questions

When should I use the ReAct prompt pattern?

You need an agent that can use tools and reason about their outputs.

What are common failure modes of ReAct?

Runaway loops • Format drift

Does ReAct work across models?

Yes when the anatomy is preserved — see model compatibility matrix on the pattern page.