What is the difference between AI Providers and AI Models? does AI Features use from AI Providers or AI Models?

[zhh210]

Having trouble understanding the usage and what is the difference between AI Providers and AI Models tab. There are quite a few overlaps between the two but I don't see which 'Model' is associated with the 'OpenAI-Compatible Provider':


Also if I were to configure my AI assistant (AI Features tab) and choosing a "Model", is it referring to the AI Providers tab or the AI Models tab? (Seems to be referring to AI Models tab but in that case from the error). I've configured an OpenAI-Compatible provider from the AI Providers tab but chat keeps giving an error of
OpenAI API key not configured. Go to Settings > AI to configure.

Also the xx AI Model doesn't use xx AI Provider as when I specify base_url and token on OpenAI provider and use OpenAI model it gives Error code: 404 - {'detail': 'Not Found'}

The AI Models tab does allow adding a model but you'll need to specify the 'Provider' which only shows the list from the AI Models tab but no 'OpenAI-Compatible' option.

[Light2Dark]

If you add a model in AI models and choose Custom provider, it will use the OpenAI-Compatible config under the AI Provider tab, regardless of the name of your Custom provider.

It does feel unclear. And maybe we should allow adding multiple custom providers and naming them, so you can easily tie to the model chosen.

[Light2Dark]

I can convert this to an issue for us to improve.