Inconsistent behavior with %%ai openai-chat:gpt-4 and %%ai openai-chat:gpt-3.5-turbo

[startakovsky]

Description

Looks like I am getting an error when using the product as directed. See annotated screenshot below indicating:

1. The failure of %%ai openai-chat:gpt-4
2. The success of %%ai openai-chat:gpt-3.5-turbo
3. The success of using the openai python library's chat.completions endpoint.

Error code: 404 - {'error': {'message': 'The model `gpt-4` does not exist or you do not have access to it. Learn more: https://help.openai.com/en/articles/7102672-how-can-i-access-gpt-4.', 'type': 'invalid_request_error', 'param': None, 'code': 'model_not_found'}}

Reproduce

[startakovsky]

Seems that this relies on this code from langchain, and it's deprecated. Wonder if that has to do with anything.

jupyter-ai/packages/jupyter-ai-magics/jupyter_ai_magics/providers.py

Lines 24 to 30 in 642ac53

	from langchain_community.chat_models import (
	AzureChatOpenAI,
	BedrockChat,
	ChatAnthropic,
	ChatOpenAI,
	QianfanChatEndpoint,
	)

Screenshots of ChatOpenAI and AzureChatOpenAI being deprecated from within the langchain_community module directory.

ChatOpenAI deprecation redirect to langchain_openai

AzureChatOpenAI deprecation to langchain_openai

[startakovsky]

Yeah . . looks like the langchain link says the alternative (according to the kwargs passed to the deprecated decorator would be to do this:

from langchain_openai.chat_models import (
    AzureChatOpenAI,
    ChatOpenAI,
)

I must say this is strange because I would think langchain would be wanting to provide that consistency to its community, but I'm sure they had a good reason.