Connection error when using other models

[InstaZDLL]

[EN] I have a problem when I want to use another model, I have this error, I changed the server url endpoint as indicated by OpenAI here.

In my example, I wanted to use text-davinci-003 but it doesn't work for all the other models except gpt4 (which I can't test), gpt3-turbo and gpt3-turbo-0301.

Even if I uncheck the box I have the same problem (My settings) : what am I doing wrong ?

[CN] third part translation tool - 第三部分翻译工具 我有这样的问题，当我想使用的另一个模型时，我有这样的错误，我改变了OpenAI所指示的服务器的URL端点这里.

在我的例子中，我想使用text-davinci-003，但除了gpt4（我无法测试）、gpt3-turbo和gpt3-turbo-0301，它对所有其他型号都不起作用。

即使我取消勾选，也有同样的问题（我的设置）：我做错了什么？

[obiscr]

Hi, @InstaZDLL

The official Server URL is: https://api.openai.com/v1/chat/completions
And you fill in is:         https://api.openai.com/v1/completions

The /chat has been lost in your configuration.

For model: If you select the gpt-4, It is only useful for accounts that have GPT-4 access. If your account does not yet have GPT-4 access, then you will not be able to use GPT-4.

You also mentioned the text-davinci-003 model, which should be caused by the official API :https://api.openai.com/v1/chat/completions not being supported.

[InstaZDLL]

Ok so we can't use text-davinci-003 with this plugin because even if I put /chat or not I get the same error even though it's well presented on the OpenAI website :

[obiscr]

Thank you for the information, I will test it further.

[obiscr]

After testing and consulting the official documentation, I have found that this is a bug. The error occurs because the data structure has changed. However, currently, apart from the gpt-4 model, gpt-3.5-turbo should be the best model available. Previous models like text-davinci-003 and text-davinci-002 were not as useful as gpt-3.5-turbo. Therefore, there are no plans to support the older models unless it is absolutely necessary.