feat: add GPT-5.6 Codex models and fix provider probes

[ascendswang]

Summary

• send ChatGPT/Codex provider probes through the Responses API in stateless streaming mode
• keep Responses API output in the current string-compatible format and preserve account headers
• add GPT-5.6 Sol, Terra, and Luna to OAuth-created providers and the built-in Codex preset
• add regression coverage for the Codex-specific request options and model catalog

This fixes provider tests failing with:

Error code: 400 - {'detail': 'Store must be set to false'}

The GPT-5.6 model IDs follow the current OpenAI model guidance.

Target branch

• Base is develop (feature / fix — default)
• Base is main (release/* or hotfix/* only)

Type of change

• Bug fix
• New feature
• Breaking change
• Documentation
• Refactor / chore
• Release / hotfix

Test plan

• make all passes locally (1531 passed, 6 skipped, 31 deselected)
• Added/updated tests
• Live Codex provider probe returns ok
• Live probes pass for gpt-5.6-sol, gpt-5.6-terra, and gpt-5.6-luna

Checklist

• Updated CHANGELOG.md (if user-facing)
• README / docs updated (not needed)