Azure OpenAI Service
An Azure service that provides access to OpenAI’s GPT-3 models with enterprise capabilities.
4,080 questions
Image Input Support for Azure OpenAI Assistant API
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(288, 41%, 37%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3ESF%3C/text%3E%3C/svg%3E)
Schmidt, Friedrich
15
Reputation points
I'm currently working with the Azure OpenAI Assistant API and have encountered a limitation. I'm trying to use image input with the GPT-4o model, but it seems this feature is not yet supported.
Details:
- Model: GPT-4o
- API Version: 2024-07-01-preview
- Error: Receive a 400 error stating only 'type=text' messages are supported
Is there an estimated timeline for when image input will be supported in the Azure OpenAI Assistant API? Any official information or guidance would be greatly appreciated. Thank you for your time and assistance.
Azure OpenAI Service
{count} votes
-
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(6.4, 61%, 10%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3ES%3C/text%3E%3C/svg%3E)
santoshkc 15,245 Reputation points Microsoft External Staff Moderator2024-09-06T15:28:51.99+00:00 Thank you for reaching out to Microsoft Q&A forum!
According to the available information, the integration with the Assistants API is not supported. For more info, please look into: Azure OpenAI global batch_API Version_Not supported.
However, I tried to repro the issue with the image input, using Global-standard API versions: 2024-08-06 and 2024-05-13.
You can keep an eye on the Azure OpenAI Service release notes for updates on What's new in Azure OpenAI Service.
I hope you this helps! Thank you.
-
Schmidt, Friedrich 15 Reputation points
2024-09-09T07:58:46.7533333+00:00 Hi @santoshkc ,
Thank you for your reply and for testing the image input with the global-standard API versions (
2024-08-06and2024-05-13). My deployments are standard deployments in Sweden Central. Here image inputs using the normal Chat Playground are working, too.However, I would like to clarify that I am specifically working with the Assistant API, not the batch API. My goal is to use image inputs with the Assistant API, where I'm currently getting the error.
It seems the Assistant API does not yet support image inputs, unlike the standard GPT API, which you referred to in your test. I was wondering if you could provide any insights into when image input support might be available for the Assistant API, or if there is a different approach or timeline for this feature?
Thank you again for your help!
-
Schmidt, Friedrich 15 Reputation points
2024-09-09T08:00:56.9+00:00 Thank you for your prompt reply and for testing the image input with the global-standard API versions (
2024-08-06and2024-05-13). However, I would like to clarify that I am specifically working with the Assistant API in Sweden Central, not the batch global API. My goal is to use image inputs with the Assistant API, where I'm currently getting the error.It seems the Assistant API does not yet support image inputs, unlike the standard GPT API, which you referred to in your test. I was wondering if you could provide any insights into when image input support might be available for the Assistant API, or if there is a different approach or timeline for this feature?
Thank you again for your help!
-
santoshkc 15,245 Reputation points Microsoft External Staff Moderator
2024-09-10T16:56:31.0466667+00:00 Hi @Schmidt, Friedrich,Thank you for the clarification. As you've noted, the Assistant API currently does not support image inputs, which explains the 400 error you're encountering. While image inputs do work through the Chat Playground using the standard GPT API, this functionality has not yet been extended to the Assistant API, even in the Sweden Central region.
At this time, there is no specific timeline available for when image input support will be introduced to the Assistant API. However, you can keep an eye on the Azure OpenAI Service release notes for any updates regarding this feature. If image input is a critical requirement for your project, using the GPT API through the Chat Playground might be the most practical solution until the Assistant API is updated to include this capability.
I hope you understand! Thank you.
Sign in to comment
Sign in to answer