.Net: Bug: Property 'max_completion_tokens' not found on type 'Microsoft.SemanticKernel.Prompty.Core.PromptyModelParameters'.

[sundium]

Describe the bug
Using Prompty module with GPT 5o models that does not support max_tokens parameter anymore. The new parameter is max_completion_token.

Property 'max_comUnhandled exception. (Line: 12, Col: 5, Idx: 290) - (Line: 12, Col: 26, Idx: 311): Property 'max_completion_tokens' not found on type 'Microsoft.SemanticKernel.Prompty.Core.PromptyModelParameters'.

To Reproduce
Steps to reproduce the behavior:

1. Specify max_completion_tokens as a model parameter in a prompty file.
   e.g.
   configuration: type: azure_openai azure_endpoint: ${env:AZURE_OPENAI_ENDPOINT} azure_deployment: ${env:AZURE_OPENAI_DEPLOYMENT} parameters: temperature: 1

2. Create function from prompty.
   `
   var kernel = Kernel.CreateBuilder()
   .AddAzureOpenAIChatCompletion(deployment, endpoint, key)
   .Build();

var function = kernel.CreateFunctionFromPromptyFile("../basic.prompty");
3. Invoke function.var result = await function.InvokeAsync(kernel, kernelArguments);`

Expected behavior
Can invoke a GPT 5o series models.

Screenshots

Platform

• Language: C#
• Source: Top-level Package
  `dotnet list package
  Requested Resolved

  Microsoft.SemanticKernel 1.22.0 1.22.0
  Microsoft.SemanticKernel.Prompty 1.23.0-alpha 1.23.0-alpha
  `

• AI model: [e.g. OpenAI:GPT-5o-mini(2024-07-18)]
• IDE: VS Code
• OS: Windows, Mac

Additional context
Add any other context about the problem here.

[rogerbarreto]

Have you tried using the SetNewMaxCompletionTokensEnabled property? Let me know if that works for you, this is a know error that can happen for Azure deployments and there's a flag that informs the SDK on what property to use in the backend.

// Initialize your existing kernelArguments providing the settings... 
AzurePromptExecutionSettings settings = new() { SetNewMaxCompletionTokensEnabled = true };
KernelArguments kernelArguments = new(settings) 
{
   // Add your arguments here...
}

var function = kernel.CreateFunctionFromPromptyFile("../basic.prompty");
3. Invoke function.var result = await function.InvokeAsync(kernel, kernelArguments);`

Please see here:

• .Net: Bug: GPT-5 max token parameter name changed #12892