Jul 21, 2025
Most founders, and people in general, use just one LLM (large language model). This is fine, it gets the job done for most people, but we can do far better. By using different LLMs for different tasks, you can work faster and get better results. Here's a quick guide:
ChatGPT:
ChatGPT is my personal favorite LLM right now, it has a decent context window, great reasoning, and the best UI (in my opinion). You should use ChatGPT for your AI assistant (check out my article on that!) and any general questions that don't need a large context window or a deep understanding of a complex subject.
Gemini:
Gemini is great, it has a very large context window and very very good deep research. Some people might argue Openai or perplexity has a better deep research function but in my opinion, they don't even come close. Perplexity lacks the same reasoning as Gemini in deep research and o1 is way too slow.
Claude:
Claude 4 Opus is a game-changer when it comes to development. It generates extremely large pieces of code with very high accuracy and is great at debugging. It has a graduate level understanding and reasoning for technical tasks. Claude is great if you have questions or want to preform tasks around physics, math, or coding.
In a world of fast-changing AI, your competitive edge isn't just using AI—it's using the right one for the job. Always research new models when they come out because they are constantly getting better and better at different things.