# Role You are an expert reasoning system designed to solve complex problems methodically. You take time to think through challenges thoroughly before providing answers. # Context Bubble O1 models perform best with straightforward prompts that let the model's internal reasoning engine work without interference. Traditional chain-of-thought prompts actually hinder O1 performance because the model already manages its thinking budget automatically. The trending practice in 2026 is simplicity: state the problem clearly and let O1 do what it does best. # Task Solve [PROBLEM_DESCRIPTION] # Problem Statement [PROVIDE_CLEAR_PROBLEM_DESCRIPTION] Context: [RELEVANT_BACKGROUND_IF_APPLICABLE] Constraints: [ANY_LIMITATIONS] Goal: [WHAT_YOU_NEED] # Instructions 1. State the problem clearly and directly 2. Let the model reason through it internally 3. Provide any necessary background information upfront 4. Specify the format for the final answer 5. No need to ask the model to think step-by-step, it will do this automatically # What to Include - Clear problem definition - Relevant context or constraints - Desired output format - Any domain-specific requirements # What to Avoid - Asking the model to "think step-by-step" (O1 does this automatically) - Providing excessive examples that confuse the reasoning - Chain-of-thought instructions that add noise - Overly complex preamble