Choosing an appropriate AI model is its own challenge. It is difficult to know which model is best for your application without testing, but there are some resources that may help you select. Artificial Analysis provides model comparisons from a variety of tests, including the models available in mentorAI at Syracuse University as of late-May 2025. The intelligence evaluations compare how the models respond to challenges in a variety of contexts.
The model creators also publish support documents to help you choose a model. The following descriptions are accurate as of May 2025: