How to use SDK to also use image as input and ask ResponseAPI to use webSearch

[devilnest]

I have this litte call

var response = await chatClient.CreateResponseAsync(messages, 
new ResponseCreationOptions
            {
                Tools = { ResponseTool.CreateWebSearchTool(WebSearchToolLocation.CreateApproximateLocation("Poland")) }
            });

How I build messages:

var contentParts = new List<ResponseContentPart>
        {
            ResponseContentPart.CreateInputTextPart("""
                                                    Mysterius prompt
                                                    """)
        };

        foreach (var filePath in Directory.GetFiles("./", "slice*.jpg"))
        {
            var fileName = Path.GetFileName(filePath);
            var file = await File.ReadAllBytesAsync(filePath);
            var imagePart = ResponseContentPart.CreateInputImagePart(
                BinaryData.FromBytes(file), "image/jpeg", fileName );

            contentParts.Add(imagePart);
        }
        var messages = new List<ResponseItem>
        {
            ResponseItem.CreateDeveloperMessageItem(contentParts)
        };

As response:
API: HTTP 400 (invalid_request_error: invalid_value)
Parameter: input[0].content[1]

Invalid value: 'input_image'. Supported values are: 'input_text'.

So... How to use this? Images should be passed as text? (Doesn't make sense)