GPT-4.1 is out and here are some tips from inside OpenAI.

Agentic workflows: This is your new starting point. Just add tools and prompt to make it plan, execute, and verify iteratively. Only add multi-agent frameworks once you've squeezed as much as you can here.

Chain of thought: Prompting for chain of thought can help to break down the problem. Start with a simple instruction at the end of your prompt.

Long context: The full context window (now 1M token input) is much more performant/usable than before.

Improving instruction following: Performance is much better. Many model performance issues actually come from inconsistent, conflicting, or underspecified instructions. Be careful with "always."

Prompt structure: Context at or near the bottom seems to be best. Markdown and XML both work great.