Skip to main content

Start here

Which model should I use?

The honest short answer, and a quiz for the real one.

The short answer

For most people, most of the time, a mid-tier model from any major company, like Claude Sonnet or GPT-5.6 Terra, is fast, cheap, and smarter than you need. Save the premium flagships for hard coding, research, and long autonomous work, and use a fast tier for bulk simple stuff.

If your question involves anything recent, like news, prices, or sports, make sure the model can search the web. Models only know what existed when their training ended.

The real answer

It depends on what you're doing, what you'll spend, and whether you care about open source. That's exactly what our quiz asks: four questions, one recommendation, reasoning included.

Next up: Claude vs GPT: Which should I use?