Does Kernel Memory not support reasoning models such as gpt-o1 on purpose?

[piffy76]

Hello,

If I want to use gpt-o1 with Semantic Kernel Chat Completion, I need to use AzureOpenAIPromptExecutionSettings to set the SetNewMaxCompletionTokensEnabled property to 'true' because gpt-o1 requires it to be set. However, when Kernel Memory sends a RAG prompt to an LLM, it uses OpenAIPromptExecutionSettings which does not support the property, therefore, I cannot use gpt-o1 with Kernel Memory. Is this limitation intentional or is it okay if I contribute to address this as an issue (trying to find out if there might be any reason not to)?