[FEATURE] Per-command model selection for Claude Code slash commands

Resolved 💬 3 comments Opened Jan 15, 2026 by Pringled Closed Jan 15, 2026

Preflight Checklist

  • [x] I have searched existing requests and this feature hasn't been requested yet
  • [x] This is a single feature request (not multiple features)

Problem Statement

Claude Code currently applies a single model selection to an entire session, unless it's manually switched. In real development workflows, task complexity varies significantly within the same session (e.g., deep architectural planning vs. straightforward implementation vs. quick review) and per command.

Proposed Solution

Allow specifying the model at the slash-command level, so a single Claude Code session can temporarily switch models for a given command invocation, then return to the session default.

Alternative Solutions

I've tried hacking around this by making aliases for starting Claude Code with a specific model but this is not a very nice solution.

Priority

Low - Nice to have

Feature Category

CLI commands and flags

Use Case Example

For example I would like to define a command like this, where I can set which model should be used for this command:

<!-- model: claude-opus-4-5-->

Plan a feature: $ARGUMENTS

1. Identify requirements and constraints
2. Propose a design and milestones
3. Call out risks and edge cases
4. Provide an implementation outline

This way, I can easily use one model as my default, and automatically use a different model when using a command like /review.

Additional Context

_No response_

View original on GitHub ↗

This issue has 3 comments on GitHub. Read the full discussion on GitHub ↗