feat: support models served via Open AI API

[sdirix]

What it does

The Open AI settings now allow to configure additional models and their associated endpoints. This can be used to integrate any custom model running locally or in the Cloud as long as it supports the Open AI API.

fixed #14174

How to test

1. Serve a model via the Open AI API. The following steps assume you run a model on Runpod with vllm
2. Open "Open AI" preferences and add a custom model, e.g.

        {
            "model": "mistralai/Mistral-7B-Instruct-v0.2",
            "url": "https://<your-pod-id>-8000.proxy.runpod.net/v1",
            "id": "my-mistral2"
        }
   

3. Configure and use model

Follow-ups

• For now we disable "tool" usage for non-Open AI models. We should allow to enable / configure tool usage in some way in the future

Review checklist

• As an author, I have thoroughly tested my changes and carefully followed the review guidelines

Reminder for reviewers

• As a reviewer, I agree to behave in accordance with the review guidelines

[planger]

Very nice, works great and code looks good! Thank you!