[FEATURE] Per-tool model routing in settings.json

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's model setting is session-wide. There's no way to configure different models for different tools within the same session. I want expensive/slow tasks (code generation, analysis) to use Sonnet, while lightweight shell/bash execution decisions use Haiku — reducing cost and latency without degrading quality where it matters.

Proposed Solution

"modelOverrides": {
"tools": {
"Bash": "haiku",
"Read": "haiku"
}
}
When a tool call is made, the matching model handles that turn's reasoning.

Priority

Medium - Would be very helpful

Feature Category

Configuration and settings

Use Case Example

_No response_

Additional Context

_No response_