docs: refresh model references to current models

[roomote]

This PR updates documentation to remove outdated references (GPT-4, GPT-4.5 Preview, Claude 3.5) from non-historical docs. Changes:

• Update VS Code LM examples to claude-3.7-sonnet.
• Update concurrent-file-edits best-practices to Claude 4/3.7, GPT-4.1/4o, and GPT-5.
• Remove gpt-4.5-preview from OpenAI docs and OpenAI-compatible model lists.
  Rationale: address public feedback indicating FAQs mention old models.

---

Important

Update documentation to reflect current AI models, removing outdated references and ensuring consistency across docs.

• Documentation Updates:
  • Update concurrent-file-edits.md to reference Claude Sonnet 4, Claude 3.7 Sonnet, GPT-4.1/GPT-4o, and GPT-5.
  • Update claude-code.md to remove Claude 3.5 Sonnet and Claude 3.5 Haiku from recommended models.
  • Update litellm.md to use claude-3.7-sonnet-20250219 for supportsComputerUse flag.
  • Remove gpt-4.5-preview from openai-compatible.md and openai.md.
  • Update vscode-lm.md to use claude-3.7-sonnet in examples.
• Rationale:
  • Address public feedback about outdated model references in FAQs and documentation.

^{This description was created by for 7275bab. You can customize this summary. It will automatically update as commits are pushed.}

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Preview	Comments	Updated (UTC)
roo-code-docs	Ready	Preview	Comment	Oct 17, 2025 7:39am

[hannesrudolph]

@roomote Please narrow this PR to focus only on recommended models, not all model mentions.

1. Scope

   • This PR should update recommended or default model references only.
   • Do not remove or edit mentions of older models that appear for historical or compatibility documentation (e.g., “previous versions,” “legacy models,” or “migration examples”).

2. Anthropic

   • Replace all recommended mentions of "Claude 3.7 Sonnet" and "Claude 4" with:
     • Claude Sonnet 4.5 (default)
     • Claude Opus 4.1 (premium reasoning)

3. OpenAI

   • Use only GPT-5 for all recommended model examples.
   • Remove GPT-4, GPT-4.1, GPT-4o, and GPT-4.5 Preview from recommended sections.
   • Keep them if referenced historically (e.g., “supported previously” or “for legacy users”).

4. Google

   • Update to Gemini 2.5 Pro and Gemini 2.5 Flash as recommended models.

5. xAI

   • Use Grok-4-Fast (2M context) and Grok-4-Fast-Reasoning for any reasoning examples.

6. General

   • Remove hardcoded suffixes (e.g., -20250219).
   • Use model family names (Claude Sonnet 4.5, GPT-5, etc.).
   • Ensure only recommended models in tables, examples, and “supported providers” lists are updated.

Summary:
This PR is for recommended model updates only. Keep legacy references where contextually relevant but update all recommended defaults to Claude Sonnet 4.5, Claude Opus 4.1, GPT-5, Gemini 2.5 Pro/Flash, and Grok-4-Fast.