- Fix typo 'Start' 'End' - Add llama3.1 selection - Use 'quick_think_llm' model instead of hard-coding GPT