Can not access to o1 for vision task

API
1

This is my code to access the api:

        const promptBody = {
            "model": "o1-2024-12-17",
            "input": [
                {
                    "role": "user",
                    "content": [
                        { "type": "input_text", "text": "" },
                        {
                            "type": "input_image",
                            "image_url": ""
                        }
                    ]
                }
            ]
        };
        if (textPrompt) {
            promptBody.input[0].content[0].text = textPrompt
        }
        if (imageBase64) {
            promptBody.input[0].content[1].image_url = imageBase64
        }

        if (messages.length === 0) {
            throw new Error('Either textPrompt or imageBase64 is required');
        }

        const payload = promptBody
        // Generate a unique request ID
        const requestId = `req_${Date.now()}_${Math.random().toString(36).substr(2, 9)}`;

        // Start progress simulation for OpenAI
        pollStatusOpenAI(clientId, requestId);

        // Send request to OpenAI
        const response = await axios.post(openaiUrl, payload, {
            headers: {
                'Authorization': `Bearer ${openaiApiKey}`,
                'Content-Type': 'application/json'
            },
            timeout: 10000 // 10-second timeout
        });

When I send the request, I receive this error message:
404 The model o1-2024-12-17 does not exist or you do not have access to it.

Is there anything wrong with my request?

2

Yes.

  1. Were you to have access to the o1 series of models, that timeout would disconnect and throw an error long before the model could return a reasoning response to you, especially with an image.

  2. The image_url field is a data URL, not simply the base64 data alone, as you can clearly see here:

import base64,openai;print(openai.Client().chat.completions.create(**{"model":"gpt-4o","messages":[{"role":"system","content":"Vision assistant"},{"role":"user","content":[{"type":"text","text":"Describe image"},*([{"type":"image_url","image_url":{"url":(f"data:image/{__import__('pathlib').Path(f).suffix[1:]};base64,{base64.b64encode(__import__('pathlib').Path(f).read_bytes()).decode()}"),"detail":"low"}}for f in["i.png"]])]}],"max_completion_tokens":100}).choices[0].message.content)

  1. There also is nothing to think hard about in an image that a model like gpt-4o-2024-11-20 can’t do at a lower price, unless it is “solve this visual puzzle that no other AI can perform well on”.

Then of course finally: why don’t you have access?

Only a subset of accounts have access. Use the API models endpoint or the Prompt Playground and see it not available there either.

https://help.openai.com/en/articles/10362446-api-access-to-o1-and-o3-mini