Cost of API. Getting chagred for GPT-3.5-turbo as well as text-davinci:003

In my billing section, I can see charges for GPT-3.5-turbo as well as Base Model (text-davinci:003). The billing for GPT-3.5-turbo seems to be ok, but that for text-davinci is very high. Why am I being charged for both and what can I do to reduce the costs? I tried initialising the model in both of the following ways

  1. llm = OpenAI()
  2. lllm = ChatOpenAI(openai_api_key = openai.api_key, model_name = “gpt-3.5-turbo-0613”, temperature = 0,) langchain

I am using Laghchain’s create_pandas_dataframe_agent to query a Pandas dataframe.

The first method indicates completions, and may have text-davinci-003 set in it as a default model.

Do a file text search on your langchain directory and change every “davinci” to “leonardo” with a replace… Then you get errors instead of calls to AI models you didn’t expect.

Langchain in general takes high understanding before letting it loose on your account balance.

The price of davinci class gpt-3 IS high.