In Claude Code, the advisor is a second model that the main session may consult for guidance at important decision points. It does not take over the coding session: the main model continues handling the task, while the advisor offers a second opinion that Claude can consider.
What the advisor does
Anthropic describes the advisor as a second, typically stronger model Claude can consult before choosing an approach, when it encounters a recurring error, or before it says a task is complete. The advisor receives the conversation, including tool calls and results, and returns guidance to the main session. It advises; it does not ordinarily carry out the task in place of the session model.
![Claude AI for Beginners Bible: [5 in 1] The Ultimate Guide to Automate Your Work, Save Hours Every Week, and Use AI for Real-World Results](https://m.media-amazon.com/images/I/415+fSJacsL._SL160_.jpg)
Claude Code decides when to make an advisor call. The timing is model-driven, not a fixed checkpoint schedule. You can ask Claude to consult the advisor in your instructions, but the documentation does not provide a setting to force or cap the number of calls. See Anthropic’s advisor documentation.
Advisor vs. main coding model vs. subagent
These features involve different roles: the advisor offers guidance to the main thread, while a subagent can take on delegated work in its own context.
The Tool Desk
Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →| Claude Code feature | Role | Does it replace the main session model? |
|---|---|---|
| Main coding model | Handles the ongoing task and responds to you. | It is the model assigned to the session. |
| Advisor | Provides guidance when Claude consults it at a decision point; it receives the conversation, including tool calls and results. | No. The main model continues the task. |
| Subagent | Handles a delegated subtask in its own context, with its own system prompt, tool access, and permissions, then returns results. | No. It performs a specific piece of work rather than advising the main thread. |
Anthropic describes subagents as specialized assistants for specific types of tasks. Details are in the subagent documentation.
How to enable or change the advisor
You can manage the advisor in three ways, depending on whether you want to change it interactively, configure it, or use it for just one session:
Rank #2
- Use
/advisorto enable or change the advisor in Claude Code. - Set
advisorModelin settings for configuration. - Pass
--advisorwhen starting a session to specify it for that session.
The advisor must meet Anthropic’s compatibility requirements for the main model. Pairings may be equal-ranked or higher-ranked; an arbitrary advisor model is not necessarily supported. Check the current compatibility details in the advisor documentation.
When an advisor is useful
An advisor is most relevant when a task has consequential decisions and benefits from an independent review, rather than requiring the strongest model on every routine turn. Anthropic suggests considering it for:
Free tools Windows power users keep installed
One-click scans. No signup required.
- Large refactors that require a sound plan.
- Debugging a recurring error.
- Reviewing an approach before a major commitment or checking work before completion.
For a short task with little planning, the additional consultation may add little. If every turn needs the strongest model, using that model as the main session model may be more appropriate than relying on occasional advisor calls.
Advisor vs. other model controls
The advisor is not the only way to change how Claude Code uses models, but the controls serve different purposes:
/modelchanges the model used for requests from that point onward; it is not an on-demand second opinion.opusplan, where supported by model settings, uses a model during plan mode and then switches to Sonnet for execution.
Anthropic’s model configuration documentation describes model selection and configuration.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Usage, availability, and limitations
Each advisor call processes the conversation with the advisor model, so it adds model usage at that model’s rates. With API billing, the advisor’s input and output tokens are billed at its rates. Subscription use generally counts against plan limits, with Fable usage credits applying on some plans. The total varies with the calls made, the model, and the billing arrangement; Anthropic’s documentation does not establish one universal cost or savings figure.
The feature is experimental, so behavior, pricing, and availability may change. It requires the Anthropic API and is not available through Amazon Bedrock, Claude Platform on AWS, Google Cloud’s Agent Platform, or Microsoft Foundry. Consult the current advisor documentation for supported configurations.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
