Models like gpt-5-search-api were incorrectly matching the gpt-5 pattern and getting temperature=1.0, when they should match the -search pattern and get nil.

Reordered conditions to check for -search models first, ensuring the more specific pattern takes precedence.

What this does

Fixes a bug in normalize_temperature where the condition order caused search models in the GPT-5 family (e.g., gpt-5-search-api) to incorrectly receive temperature=1.0 instead of nil.

The fix reorders the pattern matching so the more specific -search check happens before the broader gpt-5 check, ensuring search models don't accept the temperature parameter as required by the OpenAI API.

Type of change

• Bug fix
• New feature
• Breaking change
• Documentation
• Performance improvement

Scope check

• I read the Contributing Guide
• This aligns with RubyLLM's focus on LLM communication
• This isn't application-specific logic that belongs in user code
• This benefits most users, not just my specific use case

Quality check

• I ran overcommit --install and ⚠️ some hooks failed, but in files unrelated to the fix.
• I tested my changes thoroughly
  • For provider changes: Re-recorded VCR cassettes with bundle exec rake vcr:record[provider_name]
  • All tests pass: bundle exec rspec
• I updated documentation if needed
• I didn't modify auto-generated files manually (models.json, aliases.json)

API changes

• Breaking change
• New public methods/classes
• Changed method signatures
• No API changes