refactor(azure-open-ai): rename max_tokens to max_completion_tokens and remove deprecated parameters across multiple models

[LordGameleo]

• Updated parameter names from max_tokens to max_completion_tokens for consistency.
• Removed deprecated max_tokens parameter from various model YAML files.
• Ensured all models reflect the latest naming conventions and parameter structures.

This change enhances clarity and aligns with the updated API specifications.

---

Note

Medium Risk
Model catalog changes that alter which token-limit parameter is sent to Azure OpenAI. Incorrect mapping could cause API errors or ignored output-length settings for many GPT-5 models.

Overview
Aligns Azure OpenAI GPT-5 family model configs with the Chat Completions API by exposing max_completion_tokens instead of max_tokens.

Across GPT-5 through GPT-5.6 YAML definitions, the request param key is renamed (or added where missing), and max_tokens is listed under removeParams so the deprecated field is stripped from requests. Limits still keep max_tokens as a catalog limit field.

^{Reviewed by Cursor Bugbot for commit fbcc844. Bugbot is set up for automated code reviews on this repo. Configure here.}

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 2 potential issues.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

Want higher recall? High effort reviews run extra passes and find more bugs. A team admin can switch effort levels in the Cursor dashboard.

^{Reviewed by Cursor Bugbot for commit 612af3e. Configure here.}