gpt-5.4-mini needs Response API

[Bloody-Wolf799]

Is it a request payload issue?
[x] Yes, this is a request payload issue. I am using a client/cURL to send a request payload, but I received an unexpected error.
[ ] No, it's another issue.

If it's a request payload issue, you MUST know
Our team doesn't have any GODs or ORACLEs or MIND READERs. Please make sure to attach the request log or curl payload.

Describe the bug
GitHub Copilot gpt-5.4-mini: bad response status code 400, message: model "gpt-5.4-mini" is not accessible via the /chat/completions endpoint, body: {"error":{"message":"model "gpt-5.4-mini" is not accessible via the /chat/completions endpoint","code":"unsupported_api_for_model"}}

CLI Type
OpenAI Compatible

Model Name
gpt-5.4-mini

LLM Client
OpenCode or simple HTTP requests

Expected behavior
Like gpt-5.x-codex, auto translate to Responses API.

[hungthai1401]

#476 wait for merging