Expose model reasoning metadata for Codex

[khaliqgant]

Summary

• add reasoning-effort metadata for Codex models to the shared CLI registry
• generate config metadata helpers and re-export them from the SDK
• add SDK coverage for the Codex mini/high and frontier/xhigh defaults

Verification

• npm run codegen:models:ts
• npm --prefix packages/sdk run build
• node --input-type=module -e "import { ModelMetadata, getDefaultReasoningEffort } from './packages/config/dist/index.js'; if (getDefaultReasoningEffort('codex', 'gpt-5.1-codex-mini') !== 'high') throw new Error('mini'); if (ModelMetadata.Codex['gpt-5.4']?.defaultReasoningEffort !== 'xhigh') throw new Error('frontier'); console.log('ok');"

---

[devin-ai-integration]

✅ Devin Review: No Issues Found

Devin Review analyzed this PR and found no potential bugs to report.

View in Devin Review to see 3 additional findings.