active Thinking Server Port
The port used by the dedicated thinking LLAMA-Server instance (if configured). Used by AiProxy to route thinking-mode requests to the correct server. 0 means no dedicated thinking server is running (hybrid mode or not configured).