quick refactor to allow for easily adding new model providers. This removes our usage of the openAIChat provider.
However trying with the new anthropic models is leading to errors due to bad api endpoints being used in llm-api. @matildepark we should sync on what to do about llm-api because it is becoming a headache
quick refactor to allow for easily adding new model providers. This removes our usage of the openAIChat provider.
However trying with the new anthropic models is leading to errors due to bad api endpoints being used in
llm-api
. @matildepark we should sync on what to do aboutllm-api
because it is becoming a headache