AI Field Guide
Models

Claude

Updated 2026-08-04

Questions this answers

  • What is Claude and who makes it?
  • How is Claude different from ChatGPT?
  • Which model should I use for writing and coding?
  • Is Claude a good default?

The Fix

Claude is the family of models from Anthropic, reached through the Claude apps, the API, or tools like Claude Code. It's proprietary, so you use it over the internet rather than running it yourself. Claude has a reputation for careful writing and strong coding, which is why a lot of people building serious tools default to it.

It comes in tiers, a faster cheaper one for routine work and a heavier one for hard problems, so you can trade speed and cost against depth. Anthropic leans hard on safety and steerability, which shows up in how closely Claude sticks to a system prompt.

When to Use It

A safe default for writing and coding, and for agents where you want output that reliably follows the brief. Reach for the heavier tier on hard problems, and drop to the lighter one for routine edits to save time and money.

Like any closed model, you're sending data to Anthropic and paying per use, so weigh that for sensitive material or high volume, where an open-weight model you host might fit better.

Best Practices

Related