Claude is a family of large language models made by Anthropic, an AI safety company founded in 2021 by former OpenAI researchers. You meet Claude in three places: the Claude app (web, desktop and mobile chat), Claude Code (a terminal and IDE agent for programmers), and the Claude API that developers use to put Claude inside their own products.
Under the hood, all three run the same models. The difference is how you pay: the app is a monthly subscription with usage limits, the API is metered per token with no ceiling. That single fact drives most of the decisions in this course.
What Claude is known for
- Writing and editing. Claude's default voice is natural and it follows tone and format instructions closely. It is the model many writers reach for first.
- Long documents. Current models hold up to a million tokens of context, so a whole codebase, a contract bundle or a year of meeting notes fits in one conversation.
- Coding and agents. Claude Code and the API's tool-use features let Claude run multi-step tasks — edit files, run tests, search the web — rather than just answer questions.
- Careful behaviour. Anthropic trains Claude to say when it is unsure, to refuse clearly harmful requests, and to follow a published set of principles. In practice this means fewer confident wrong answers, at the cost of the occasional over-cautious refusal.
What Claude is not
Claude is not a search engine. Unless you enable web search, it answers from training data with a cutoff and can be out of date. It does not generate images. And it does not remember your previous chats unless you use the app's projects and memory features, or resend the history yourself through the API.
What you are actually paying for
Every Claude product is priced on one of two meters. The app plans — Free, Pro, Max, Team, Enterprise — charge a flat monthly fee for a usage allowance measured in messages per rolling window. The API charges per million tokens in and per million tokens out, with different rates per model. At the time of writing, Pro is $20 a month and Max starts at $100 a month; API rates run from $1 per million input tokens on Haiku 4.5 to $10 on Fable 5.1.
The rest of this course is about matching those meters to your real usage. Most people who feel they are overpaying for Claude are on the wrong meter, not the wrong model. See how the plans compare in Claude Free vs paid, or jump to lesson 3.
Core tools for this lesson
Knowledge check
Which statement about Claude is true?