ChatGPT-5 from OpenAI boosts performance in math, coding, and long document processing. However, it struggles with vague prompts. To improve output quality, the article outlines four practical techniques. First, adding specific keyword phrases like "Think hard about this" or "Think deeply about this" at the end of your prompt can trigger a more advanced processing model, leading to better results. Users can also specify which model they want to use from the contextual menu, which is helpful if most tasks require detailed analysis. Second, controlling the verbosity of responses is essential. By instructing ChatGPT to aim for a certain word count or content length, users can get concise outputs tailored to their needs. For instance, asking for "3 sentences for top-level executives" or "500 words" yields precise results. Lastly, using XML format for prompts offers advantages over markdown. The structure of XML allows for clearer organization of the prompt into distinct sections, making it easier for the AI to understand context and follow instructions logically. This structured approach can enhance the quality and relevance of the AI's responses.