Feature Request / Question: Support for reasoning_effort with Ollama provider

​Context
I am using @copilot-kit/sdk, and I want to configure Ollama as my LLM provider.

​The Issue
I would like to know if it is possible to configure the reasoning_effort parameter (low, medium, high) through the SDK when using the Ollama provider (the "think" parameter in ollama takes only 2 values (True/False)) .
​Currently, I cannot find a clear way to pass this specific parameter in the provider configuration or the request options.

​Proposed / Asked behavior
Is there an existing way to pass provider-specific parameters like reasoning_effort to Ollama? If not, are there plans to support this to allow better control over "thinking" tokens?

Start by tracing the Ollama provider configuration and request options in the SDK, then compare how provider-specific parameters are handled elsewhere. Determine whether reasoning_effort or Ollama's think setting has an existing entry point, and define completion as a documented or implemented way to control it through the provider.

Written by the indexing model from the issue text.