Thinking
Ollama introduces 'Thinking' feature, allowing users to enable/disable model's thinking process for tailored application behavior. Supports CLI, API, and libraries with options to separate thinking from output or bypass it for faster responses.