Cannot use GPT-4o-mini (404 error)

API
1

I’ve paid for 5 dollars, and I’m teir 1.
I also go to the platform> settings > project >limits and the gpt-4o-mini is allowed.
I’ve copy the module name from here

here is my code

client = OpenAI(api_key=os.environ['OPEN_KEY'])

@main_bot.command()
async def ask(ctx, *, question: str):
    try:
        response = client.completions.create(
            model="gpt-4o-mini",  # Ensure you use 'model' and specify the correct model name
            prompt=question,
            max_tokens=150
        )
        answer = response.choices[0].text.strip()
        await ctx.send(answer)

    except openai.APIConnectionError as e:
        await ctx.send("The server could not be reached")
        await ctx.send(e.__cause__)  # an underlying Exception, likely raised within httpx.
    except openai.RateLimitError as e:
        await ctx.send("A 429 status code was received; we should back off a bit.")
    except openai.AuthenticationError as e:
        await ctx.send(f"Authentication failed: {e}")
    except openai.APIStatusError as e:
        await ctx.send("Another non-200-range status code was received")
        await ctx.send(e.status_code)
        await ctx.send(e.response)
    except Exception as e:
        await ctx.send(f"An error occurred: {e}")

here is my output
image

I’ve been try for the module name"gpt-4o-mini-2024-07-18",“gpt-4o-mini”,“gpt-4-turbo”,“gpt-3.5-turbo”, and “gpt-3.5-turbo-instruct”. But only gpt-3.5-turbo-instruct work, other module return the same 404 error like above output.

I also renew the API key and make sure the env key is the same as my new key, but it still not working.

How can I use gpt-4o-mini?

2

Welcome to the Forum!

The code you’ve posted will only work for instruct models such as gpt-3.5-turbo-instruct.

gpt-4o-mini and the other models you’ve listed are chat completion models. The basic API request for chat completion models looks like this:

completion = client.chat.completions.create(
  model="gpt-4o",
  messages=[
    {"role": "system", "content": "You are a helpful assistant."},
    {"role": "user", "content": "Hello!"}
  ]
)

print(completion.choices[0].message)

You can refer to the API specs for further details: https://platform.openai.com/docs/api-reference/chat

1 Like
3

Thanks for the reply!

I try to change the code into this

@main_bot.command()
async def ask(ctx, *, question: str):
    try:
        client.chat.completions.create(
          model="gpt-4o",
          messages=[
            {"role": "system", "content": "You are a helpful assistant."},
            {"role": "user", "content": f"{question}"}
          ]
        )

        await ctx.send(completions.choices[0].message)

But now it say 403 Forbidden.

4

A 403 error typically means that you are accessing the API from an unsupported country. You can check the list of supported countries here: https://platform.openai.com/docs/supported-countries

5

I pass through the 403error, but it come with the error An error occurred: 'Completions' object has no attribute 'choices'

6

My country is in the list, I’m not sure why I get the 403 error, but after I renew my API key, it didn’t happen anymore.

7

I change the completions.choices[0].message into response.choices[0].message.content.strip(), and it works!

here is the function code

@main_bot.command()
async def ask(ctx, *, question: str):
    try:
        response = client.chat.completions.create(
          model="gpt-4o-mini-2024-07-18",
          messages=[
            {"role": "system", "content": "You are a helpful assistant."},
            {"role": "user", "content": question}
          ]
        )

        answer = response.choices[0].message.content.strip()#completions.choices[0].message

        await ctx.send(answer)
    except openai.APIConnectionError as e:
        await ctx.send("The server could not be reached")
        await ctx.send(e.__cause__)  # an underlying Exception, likely raised within httpx.
    except openai.RateLimitError as e:
        await ctx.send("A 429 status code was received; we should back off a bit.")
    except openai.AuthenticationError as e:
        await ctx.send(f"Authentication failed: {e}")
    except openai.APIStatusError as e:
        await ctx.send("Another non-200-range status code was received")
        await ctx.send(e.status_code)
        await ctx.send(e.response)
    except Exception as e:
        await ctx.send(f"An error occurred: {e}")
1 Like