Mistral with Images raises an `Internal Server Error`

### Operating System

Windows

### Version Information

azure-ai-inference ==1.0.0b9
azure-core==1.34.0

### Steps to reproduce

Execute this this [notebook `Example (with image inputs)`](https://github.com/Azure/azureml-examples/blob/main/sdk/python/foundation-models/mistral/mistralai.ipynb).

### Expected behavior

I expect an LLM response describing the image.

### Actual behavior

When trying to reproduce the example with images in this [notebook](https://github.com/Azure/azureml-examples/blob/main/sdk/python/foundation-models/mistral/mistralai.ipynb), the `resp.json()` object is None and the headers indicate an `model_error`. I can successfully process images, if I send base64 encoded images directly through the image_url field.

### Addition information

_No response_

Start by running the `Example (with image inputs)` notebook linked in the issue with azure-ai-inference 1.0.0b9, then inspect the response around `resp.json()` and the `model_error` headers. Compare that path with the working base64 `image_url` input; done means the notebook returns an LLM response describing the image without an internal server error.

Written by the indexing model from the issue text.