/model opusplan is what you're looking for in Claude Code.

If you're in Claude Code and you type the /model command, you're given these options typically:

1. Default (recommended) Opus 4.6 · Most capable for complex work

2. Opus (1M context) Opus 4.6 with 1M context · Billed as extra usage · $10/$37.50 per Mtok

3. Sonnet ✔ Sonnet 4.6 · Best for everyday tasks 4. Sonnet (1M context) Sonnet 4.6 with 1M context · Billed as extra usage · $6/$22.50 per Mtok

4. Haiku Haiku 4.5 · Fastest for quick answers

Like a good restaurant, there's a secret menu item, a model called opusplan. What does it do? opusplan uses the smartest model available, Opus 4.6, when you have to do some big thinking using Plan mode (Shift + tab to enable). Once you've built a plan and Claude Code says it's ready to implement, it will exit plan mode. When opusplan is enabled, it will automatically switch to Sonnnet 4.6, a model that is faster and cheaper.

Why do this? Sonnet consumes much less of your weekly quota on a subscription plan and consumes less money on the API plan. Instead of you having to manually do this, it will handle the model switch automatically. This saves you either quota or money.

But doesn't that reduce the quality of the outcome? Not at all. Think of it like cooking. You use the smartest model to write the recipe, like you would use your most talented chef. But once the recipe is written and clear, you don't need your star chef chopping the carrots. You can delegate that to a sous chef. That's what opusplan does - once you use the smartest model to build the plan, Claude Code will hand off a highly detailed plan to the sous chef to execute it.

This is my favorite way to use Claude Code because it uses the strengths of each model, saving quota usage for the rest of the Trust Insights team while still letting me output the highest quality work.

#AI #GenerativeAI #GenAI #Claude #ArtificialIntelligence #LargeLanguageModels #MachineLearning #IntelligenceRevolution