APIRemovedInV1 Error When Using openai.ChatCompletion.create in Clean Environments (v1.45.0+ and Docker)

[rshambaugh]

Confirm this is an issue with the Python library and not an underlying OpenAI API

• This is an issue with the Python library

Describe the bug

I am encountering an APIRemovedInV1 error when using the openai.ChatCompletion.create() method in both local and Docker environments, despite using the latest versions of the OpenAI Python library (v1.45.0 and v1.46.0).

The error persists even after following the official migration guide and ensuring that the correct API is being used. I’ve tested this in multiple clean environments, including an isolated Docker container, but the issue still occurs.

The error message is:

APIRemovedInV1: You tried to access openai.ChatCompletion, but this is no longer supported in openai>=1.0.0

Expected Behavior:
I expect openai.ChatCompletion.create() to work correctly without triggering the APIRemovedInV1 error, as per the latest API documentation.

Environment:
Python version: 3.10
OpenAI Python library version: 1.45.0 and 1.46.0 (tried both)
OS: Ubuntu (both local and Docker environments)
Additional Context:
I've purged all old versions of the OpenAI library and even rebuilt environments without caching (including Docker).
The issue persists even in fresh environments with no previous configurations.
I followed the official migration guide, and the method openai.ChatCompletion.create() is being used as described in the documentation.

To Reproduce

Install openai via pip (pip install openai==1.45.0 or pip install openai==1.46.0).
Attempt to use openai.ChatCompletion.create() with the code snippet below:

Code snippets

import openai
openai.api_key = 'my api key from .env file'

response = openai.ChatCompletion.create(
    model="gpt-3.5-turbo",
    messages=[
        {"role": "system", "content": "You are a helpful assistant."},
        {"role": "user", "content": "Hello, ChatGPT!"}
    ]
)
print(response['choices'][0]['message']['content'])

OS

Windows with WSL / Ubuntu

Python version

Python 3.12.12

Library version

openai 1.45.0 & openai 1.46.0

[RobertCraigie]

I expect openai.ChatCompletion.create() to work correctly without triggering the APIRemovedInV1 error, as per the latest API documentation.

This is not supported anymore, what documentation are you referring to? Maybe something is still out of date

[rshambaugh]

Thank you for your response. I was following what I believed was the latest documentation referring to openai.ChatCompletion.create() for chat-based completions like GPT-3.5-turbo.

Can you please point me to the correct, up-to-date method for using the chat models in Python, as well as the proper documentation? I want to make sure I'm following the current best practices.

[RobertCraigie]

Here's the readme for this SDK and the OpenAI API docs which should include all the information you need!