Chat Heading issue

[1orZero]

Hi @olimorris , I just found this awesome plugin in Reddit! I like how this plugin already implemented so many Cursor (vscode fork editor) liked features. Especially the /buffers commands to easily insert any opened buffers in the chat! Its so useful!

But I have found that when I use a custom model with openai adapter, its has different behaviour. Is this normal?

Using a custom model name does not insert a new heading for the response.
https://github.com/user-attachments/assets/78fc0a20-2393-4911-91b0-bc3ca80eb0bb

Using gpt-4o inserted a heading before and after the response.
https://github.com/user-attachments/assets/0934fc69-1ef7-47a6-b566-1e6b8353fbcd

[1orZero]

OpenAi's official model name

Custom model name

[olimorris]

Thanks for raising. I can't think what this could be given the endpoints are the same.

Can you raise a bug report and share the log file? Testing for just the gpt-4o-plus

[1orZero]

I have just created an issue with the minimal config. Thanks again for making this great plugin!

Please let me know if you need a custom endpoint & a custom model to test this.
I can DM you the credentials