Receieved email for GPT4 access but API says "The model: `gpt-4` does not exist"

API
1

I received the below email saying I have GPT-4 access, however I cannot find the model in the playground or API, and I receieve an error that model gpt-4 is not found trying to use the Python OpenAI library and through raw POST requests. I am using the ChatCompletion endpoint

image
image1002×569 31.7 KB

Has anyone faced something similar? I see a deleted post from about an hour ago with the same issue

1 Like
2

I also received the invitation, changed the model to the new one, and when I asked which model it was using, it indicated that it was using the GPT-3 model. I tried creating a new API Key, but it does not recognize the new keys

3

Congratulations and welcome to the forum.

You can find out more here…

It should be available in your Playground, though, under CHAT completions. If not, give it a bit of time, I imagine.

Happy coding!

4

Took me a while to figure out a way to get it working. Feel free to provide the code your’re using, otherwise you can try generating a new API key (one created after you got access) to access GPT-4 with API. I don’t have a Python example but here is a Javascript one that is similar:

const { Configuration, OpenAIApi } = require("openai");


  const configuration = new Configuration({
    apiKey: "<INSERT-API-KEY-HERE>",
  });

  const openai = new OpenAIApi(configuration);

  const response = await openai.createChatCompletion({
    model: "gpt-4",
    messages: [
      { role: "system", content: "You are a helpful assistant." },
      { role: "user", content: "Who won the world series in 2020?" },
      { role: "assistant", content: "The Los Angeles Dodgers." },
      { role: "user", content: "Where was it played?" },
    ],
    max_tokens: 50,
    n: 1,
    stop: null,
    temperature: 1,
  });

  console.log(response.data.choices[0].message.content);

Made a quick guide here with all the details as well of how to do the request with Next.js - might be worth a look.

5

Hey folks, this should be resolved now, it was likely due to being part of multiple orgs and sending a request without specifying the OrgID in the header. Please do reach out if you run into issues like this again.