Suddenly getting `APIStatusError: Error code: 431` when trying to pass image to gpt-4o

Question

Suddenly getting `APIStatusError: Error code: 431` when trying to pass image to gpt-4o

Marc RP 0

Hi,

I suspect I am hitting some kind of rate limit. This simple example worked perfectly the first few times I run it, but after some time (I was iterating on other parts of my code) I start to get this error message whenever I try to include and image in the content

APIStatusError: Error code: 431

even though nothing about the environment has changed. I have reproduced this simple example:

def gpt4o_url(client, image_url: str, prompt: str = "What is in the image?"): 
""" Gpt-4o model using image url """ 
response = client.chat.completions.create( 
model=os.environ.get("AZURE_OPENAI_GPT4O_DEPLOYMENT")
, messages=[ { "role": "system", "content": "You are a helpful assistant to analyse images.", }, { "role": "user", "content": [ 
{"type": "text", "text": prompt}
, {"type": "image_url", "image_url": {"url": image_url}}, ], }, ]
, max_tokens=2000, temperature=0.0, )
return response


image_url = "https://github.com/retkowsky/images/blob/master/jo.png?raw=true"

result = gpt4o_url(client, image_url, "Analyse this image")

print(result.choices[0].message.content)

The code will only run if I comment out the image attachment part

# , {"type": "image_url", "image_url": {"url": image_url}}

SRILAKSHMI C 8,695 Reputation points Microsoft External Staff Moderator

2024-11-01T15:52:05.16+00:00
Hello Marc RP,

Welcome to Microsoft Q&A! Thanks for posting the question.

The APIStatusError: Error code: 431 typically indicates that the request header fields are too large. This can happen if there are too many headers or if the headers themselves are too large. Here are some steps to troubleshoot and resolve this issue:

Ensure that you are not sending unnecessary headers in your request. Sometimes, additional headers can cause the request to exceed the size limit.

You can use tools like Postman or your code’s debugging features to inspect the headers being sent.

Try encoding the images in base64 format and include them in the request body. This can help avoid issues with large headers.
{ "type": "image_url", "image_url": "<your_image_url>" }
Marc RP 0 Reputation points

2024-11-01T16:52:46.5033333+00:00

@Anonymous I am not sending anything unnecessary. Base64 format images are the same. Again, the code worked initially then suddenly stopped working (exact same code and model deployment).
Sanika Bapat 5 Reputation points

2024-11-01T17:11:39.7633333+00:00

I am also running into the same issue while using langchain and passing images. Images are being encoded base64 and the headers only have the base necessary information. The pipeline was working fine and then started returning - HTTP Request: POST https://openai/deployments/gpt-4o/chat/completions?api-version=2024-02-01 "HTTP/1.1 431 Request Header Fields Too Large"
lionel martis 5 Reputation points

2024-11-03T01:18:31.7166667+00:00

Works for gpt4 turbo but not for gpt 4o in eastus
SRILAKSHMI C 8,695 Reputation points Microsoft External Staff Moderator

2024-11-06T09:13:10.4966667+00:00
Hi Marc RP,

I recommend reporting this issue to the Azure support team. They will be able to investigate the issue further and provide a more targeted solution. You can report the issue by following these steps:

Go to the Azure portal and navigate to your OpenAI Service resource.

Click on the "Support + troubleshooting" tab.

Fill out the required information, including a detailed description of the issue and any steps you have taken to troubleshoot it.

Submit the support request.

The Azure support team will review your request and provide assistance as soon as possible Azure support.

1 answer

Your answer

SRILAKSHMI C 8,695 Reputation points Microsoft External Staff Moderator

2024-11-01T15:52:05.16+00:00

Hello Marc RP,

Welcome to Microsoft Q&A! Thanks for posting the question.

The APIStatusError: Error code: 431 typically indicates that the request header fields are too large. This can happen if there are too many headers or if the headers themselves are too large. Here are some steps to troubleshoot and resolve this issue:

Ensure that you are not sending unnecessary headers in your request. Sometimes, additional headers can cause the request to exceed the size limit.

You can use tools like Postman or your code’s debugging features to inspect the headers being sent.

Try encoding the images in base64 format and include them in the request body. This can help avoid issues with large headers.
{ "type": "image_url", "image_url": "<your_image_url>" }
Marc RP 0 Reputation points

2024-11-01T16:52:46.5033333+00:00

@Anonymous I am not sending anything unnecessary. Base64 format images are the same. Again, the code worked initially then suddenly stopped working (exact same code and model deployment).
Sanika Bapat 5 Reputation points

2024-11-01T17:11:39.7633333+00:00

I am also running into the same issue while using langchain and passing images. Images are being encoded base64 and the headers only have the base necessary information. The pipeline was working fine and then started returning - HTTP Request: POST https://openai/deployments/gpt-4o/chat/completions?api-version=2024-02-01 "HTTP/1.1 431 Request Header Fields Too Large"
lionel martis 5 Reputation points

2024-11-03T01:18:31.7166667+00:00

Works for gpt4 turbo but not for gpt 4o in eastus
SRILAKSHMI C 8,695 Reputation points Microsoft External Staff Moderator

2024-11-06T09:13:10.4966667+00:00

Hi Marc RP,

I recommend reporting this issue to the Azure support team. They will be able to investigate the issue further and provide a more targeted solution. You can report the issue by following these steps:

Go to the Azure portal and navigate to your OpenAI Service resource.

Click on the "Support + troubleshooting" tab.

Fill out the required information, including a detailed description of the issue and any steps you have taken to troubleshoot it.

Submit the support request.

The Azure support team will review your request and provide assistance as soon as possible Azure support.

Answer 1

lionel martis 5

Seems like Microsoft APIM is causing the problem. This issue had previously occurred in May and then did a hot fix in all regions. Facing the same problem again

Pootipedi, Mallikarjun 0 Reputation points

2024-11-02T11:17:29.9333333+00:00

Microsoft APIM is causing the issue. Any request that is sent directly to Azure Open AI endpoint is processed without any issues, same request sent via APIM is failing with Error code: 431

Share via

Suddenly getting `APIStatusError: Error code: 431` when trying to pass image to gpt-4o

1 answer

Your answer