- Wait for Ollama support `enable_thinking` flag to disable thinking model (see: https://github.com/ollama/ollama/issues/10492)