Feeling overwhelmed by the multitude of LLMs out there? Trust me, you're not alone. Faced with choices from giants like OpenAI, Google, and Anthropic, plus a heap of open-source options, you don’t really know where to start. Worse still you’re concerned about the restrictions that legal and your data security teams will put on your decision.
Fortunately, there is an easy way to make this decision, and I address it in Episode 7 of our AI Strategy series, How to Pick the Best LLM for Your AI Project.
You need to consider three critical constraints when making your choice: intelligence, speed, and context window size. For the overwhelming majority of projects, the answer is simple—prioritize intelligence. Start the most intelligent model available, and for most projects that is the GPT-4 model from OpenAI. Why intelligence? Because it lets you build more reliable and cost-efficient solutions, quickly.
By the time you get a real solution working - which is going to take longer and be harder than you expect - smarter and faster models will be available. GPT-4 may even be obsolete. By starting with the best, you're future-proofing your projects and focusing on challenges that really matter—like team skills, evaluation frameworks, development environments, and company education.