fix params

[harshiv-26]

Note

Low Risk
Provider metadata-only changes with no runtime code; misconfiguration could affect how max output tokens are passed to Azure for these four models.

Overview
Updates four Azure Open AI gpt-5-chat model definitions so output length is configured via max_completion_tokens instead of max_tokens, matching newer Azure/OpenAI chat APIs and the pattern already used on models like gpt-4.1-mini.

Each touched YAML swaps the params key from max_tokens to max_completion_tokens (defaults and max/min bounds unchanged where they existed). removeParams now lists max_tokens so upstream clients strip the legacy parameter—new for gpt-5-chat, gpt-5-chat-2025-08-07, and gpt-5-chat-2025-08-15; gpt-5-chat-2025-10-03 adds max_tokens to its existing removeParams block (alongside reasoning_effort).

^{Reviewed by Cursor Bugbot for commit 1e71dbc. 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 1 potential issue.

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

^{Reviewed by Cursor Bugbot for commit 1e71dbc. Configure here.}