---
title: "An agent and a model are different things"
description: "An agent is a task-working process or role; a model is the language model it uses."
slug: "agent-vs-model"
category: "agent-management"
language: en
editorial_status: published
capability_status: guide
reviewed_at: 2026-10-13
keywords: ["agent vs LLM","multi agent management","AI model switching"]
---

# An agent and a model are different things

An agent is a task-working process or role; a model is the language model it uses. More agents do not automatically mean more models, and task visibility is different from model routing.

## Why this matters for CueSpec
The current Mac companion observes Codex tasks and lets you reply in their context. A list of several tasks is not proof that CueSpec created the agents, assigned models or coordinates their work automatically.

Future multi-agent and multi-LLM ideas need separate scope: supported providers, task identity, assignment, cost visibility and authorization. A model selector alone would not solve task coordination; an agent dashboard alone would not establish provider compatibility.

## Use current controls honestly
Choose the intended conversation and use the controls actually available in its host environment. Do not assume CueSpec automatically selects a better model when a task becomes difficult or busy.

For a workflow discussion, state whether you need to see progress, send a reply, create new work, change a model or coordinate dependencies. Those are distinct capabilities and should have distinct documentation. This separation keeps software positioning clear while leaving room for the broader product direction.

## Related reading

- [The boundary of today’s agent-management features](/multi-agent-future/)
- [Multi-LLM and broader agent support](/multi-llm-roadmap/)
- [What waiting to retry means](/retrying-status/)
- [CueSpec product facts at a glance](/product-facts/)