Call from Main?

[JohnStraumann66]

When I try to specify a model: var varResponse = await api.Completions.CreateCompletionAsync(strPrompt, model: "gpt-3.5-turbo", max_tokens: maxTokens, temperature: itemp); or var varResponse = await api.Completions.CreateCompletionAsync(strPrompt, model: "gpt-4", max_tokens: maxTokens, temperature: itemp);

I get a response error:

An error occurred: Error at completions (https://api.openai.com/v1/completions) with HTTP status code: NotFound. Content: { "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": null } }

any tip on how to fix that?

[bawkee]

That is returned by OpenAI and it tells you that the models you specified (gpt3.5-turbo and gpt-4) are not compatible with the v1/completions endpoint. On their compatibility table they are, but on the playground, you can't do it. So I assume it is deprecated now, and I can't see the pricing for it either any more. The whole completion endpoint will become deprecated soon so you will have to use chat endpoint anyway.

[OkGoDoIt]

Hey @JohnStraumann66, sorry for the delayed reply. The Completions endpoint is legacy and not compatible with most of the models. Use the Chat endpoint instead. See https://github.com/OkGoDoIt/OpenAI-API-dotnet?tab=readme-ov-file#chat-api for details, but in your example:

var varResponse = await api.Chat.CreateChatCompletionAsync(new ChatRequest()
{
    Model = Model.GPT4,
    Temperature = itemp,
    MaxTokens = maxTokens,
    Messages = new ChatMessage[] {
        new ChatMessage(ChatMessageRole.User, strPrompt)
    }
})

If you really do want to use the legacy Completions API, then try the Model.ChatGPTTurboInstruct model. For example:

var varResponse = await api.Completions.CreateCompletionAsync(strPrompt, model: Model.ChatGPTTurboInstruct, max_tokens: maxTokens, temperature: itemp);