How to Choose the Right AI Models For Your Use-Case?

Enterprise AI Field Notes by ​Aish​ & ​Kiriti​

If we had a dollar for every time someone asked this question, and every time someone answered it in a one-dimensional way, we’d have enough to train our own foundation model.

Trying to find the “best” model isn’t a single-slider decision. It never is. The choice is rarely clean. In practice, it’s multi-dimensional, and you’ll have to revisit it multiple times throughout the product lifecycle. Let’s say you have all these options on a popular AI model leaderboard like LMArena, how do you choose?

We break it down into three dimensions, with a few levers in each.