Error on using completions?

[denisj]

Describe the bug
Don't know if it's a bug, but I'm trying to make Sentiment Analysis using OpenAI GPT-4. I was trying to do so using the completions API with the example of the documentation.

response = client.completions(
   parameters: {
     model: "gpt-4o",
     prompt: "Once upon a time",
     max_tokens: 5
   }
 )

But I get this error:

OpenAI HTTP Error (spotted in ruby-openai 7.3.1): {"error"=>{"message"=>"This is a chat model and not supported in the v1/completions endpoint. Did you mean to use v1/chat/completions?", "type"=>"invalid_request_error", "param"=>"model", "code"=>nil}}

With GPT3.5 too

response = client.chat.completions(
   parameters: {
     model: "gpt-3.5-turbo-0125",
     prompt: "Once upon a time",
     max_tokens: 5
   }
)

OpenAI HTTP Error (spotted in ruby-openai 7.3.1): {"error"=>{"message"=>"you must provide a model parameter", "type"=>"invalid_request_error", "param"=>nil, "code"=>nil}}

To Reproduce
Steps to reproduce the behavior:

1. Enter in a Ruby console
2. Use the code above

Expected behavior
Just receive the response from Chat GPT.

Desktop (please complete the following information):

• OS: Ubuntu 20.04.6 LTS

Additional context

• Ruby: 3.3.6
• Rails: 7.2.2
• Ruby-OpenAI: 7.3.1

[eltoob]

max token is weirdly low, did you try increasing that?

[denisj]

Still facing the same issue.

Could it be a problem on the key used that doesn't have the proper credentials?

[posiczko]

The completions models are now considered legacy. The /v1/completions endpoint is still present, however, you must use a compatible model to access them:

• gpt-3.5-turbo-instruct
• babbage-002
• davinci-002