Azure GPT 4 models need adjusted default_max_tokens

[cbagwellfnc]

Describe the bug

gpt-4o, gpt-4o-mini, and gpt-4.1 all have 16K max output tokens but json lists as 20K. Attempting to use results in this error response:

400 Bad Request { "message": "max_tokens is too large: 20000. This model supports at most 16384 completion tokens, whereas you provided 20000.", "type": "invalid_request_error", "param": "max_tokens", "code": "invalid_value" }

Changes made to openai.json in commit 530952c need to be ported to Azure to resolve the issue.

To Reproduce

Steps to reproduce the behavior:

1. Configure Azure provider
2. Select gpt-4o model
3. Send prompt
4. See error