feat(acp): move reasoning level to a separate switcher

[oskarraszkiewicz]

Issue for this PR

Closes #16543

Type of change

• Bug fix
• New feature
• Refactor / code improvement
• Documentation

What does this PR do?

This PR moves the reasoning level/model variants to a separate switcher, so the model selector is easier to use in IDEs which use ACP (Zed, JetBrains, neovim, etc.). Currently, every model variants appears as a separate model, which results in hard to use switcher. GPT-5.4 is 6 separate models right now.

After the change, it's just one model with a separate switcher.

I based the changes on how zed-industries/codex-acp handles it.

How did you verify your code works?

I tested the changes by adding a custom ACP in Zed 0.229.0 (macOS 26.4, brew install). I checked it on an actual codebase and compared the none, default and xhigh variants to compare how they work. It clearly has a different reasoning budget.

Also, I wrote test cases for it.

Didn't test in other editors.

Screenshots / recordings

If this is a UI change, please include a screenshot or recording.

(before)

(after)

Checklist

• I have tested my changes locally
• I have not included unrelated changes in this PR

If you do not follow this template your PR will be automatically rejected.

[github-actions]

Hey! Your PR title acp: move reasoning level to a separate switcher doesn't follow conventional commit format.

Please update it to start with one of:

• feat: or feat(scope): new feature
• fix: or fix(scope): bug fix
• docs: or docs(scope): documentation changes
• chore: or chore(scope): maintenance tasks
• refactor: or refactor(scope): code refactoring
• test: or test(scope): adding or updating tests

Where scope is the package name (e.g., app, desktop, opencode).

See CONTRIBUTING.md for details.

[github-actions]

Thanks for updating your PR! It now meets our contributing guidelines. 👍