I received feedback from a user that has access to a non-public model version but needs help to use it with this extension, as the list of models is currently hard-coded.
To improve upon the current behaviour, it would be sufficient to call the appropriate ListModels endpoints and to include those with the "owned_by" attribute set to "openai-internal" and "openai-dev". The challenge here is that there no way to distinguish TextCompletion from ChatCompletion models - but relying on the "gpt" prefix might be enough for now.
The improvement will be implemented in the next release version.
I received feedback from a user that has access to a non-public model version but needs help to use it with this extension, as the list of models is currently hard-coded.
To improve upon the current behaviour, it would be sufficient to call the appropriate ListModels endpoints and to include those with the "owned_by" attribute set to "openai-internal" and "openai-dev". The challenge here is that there no way to distinguish TextCompletion from ChatCompletion models - but relying on the "gpt" prefix might be enough for now.
The improvement will be implemented in the next release version.