[FEATURE] Auto-route simple tasks to Sonnet to conserve Opus quota

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

When using Claude Code, it would be useful if the system could automatically detect when a task is simple enough for Sonnet (file reads, straightforward edits, searches) and route it there instead of using the Opus quota. This would let users get more out of their "All models" allowance by reserving it for complex reasoning tasks.

Proposed Solution

Add an "auto" model mode (alongside the current manual toggle) that evaluates the incoming task complexity before selecting a model. For example:

• Route to Sonnet: file reads, grep/glob searches, simple find-and-replace edits, straightforward single-file changes, informational questions
• Route to Opus: multi-file architectural changes, complex debugging, tasks requiring deep reasoning across interconnected systems, planning/design work

The routing could be a lightweight classifier based on the user's message and current context. Users would opt in via a setting like model_selection: auto in their config, with the ability to override per-message if needed.

Alternative Solutions

_No response_

Priority

High - Significant impact on productivity

Feature Category

CLI commands and flags

Use Case Example

_No response_

Additional Context

_No response_