# Claude

> Anthropic's family of AI models, and the assistant app built on them.

- Level: Basic
- Topics: AI, Tech companies
- URL: https://tidbit.legios.com.ar/claude/
- In Spanish: [Claude](https://tidbit.legios.com.ar/es/claude/)

**Claude** is the name of the large language models made by Anthropic, and of the assistant people use to talk to them. Like other chat assistants, it answers questions, writes and edits text, summarizes long documents, analyzes data, reads images and PDFs, and writes code. The first version came out in March 2023, and Anthropic releases new generations regularly, numbered like Claude 3 or Claude 4.

The models come in sizes with different trade-offs. **Haiku** is the small, fast and cheap one; **Sonnet** sits in the middle; **Opus** is the largest and most capable, and also the slowest and most expensive. There are several ways to reach them: the Claude app on the web, desktop and phone; an API that developers pay for per token; cloud platforms such as Amazon Bedrock and Google Cloud; and **Claude Code**, a tool that lets Claude work inside a software project, reading files, changing code and running commands.

What sets it apart is mostly how it is trained. Anthropic uses a method it calls **Constitutional AI**, where the model learns to follow a written set of principles, and it publishes research on safety and on understanding what happens inside the model. In practice Claude is widely used for writing, for analysis of long documents and for programming. Like any language model, it can be wrong while sounding sure, so it is a collaborator to check, not an oracle.

## An analogy

Claude is like a firm of assistants with the same training but different seniority: a quick junior for simple errands, a reliable mid-level for most work, and a senior partner for the hardest problems, who takes longer and bills more.

## Keep in mind

- Made by Anthropic; the first version came out in March 2023.
- Three sizes: Haiku (fast), Sonnet (balanced), Opus (most capable).
- Available as an app, an API, through clouds and as Claude Code.
- Trained with Constitutional AI, guided by written principles.

## Go deeper

- [Claude on Wikipedia](https://en.wikipedia.org/wiki/Claude_(language_model))
- [Intro to Claude (official docs)](https://platform.claude.com/docs/en/intro)
- [Meet Claude (Anthropic)](https://www.anthropic.com/claude)

## Related terms

- [Anthropic](https://tidbit.legios.com.ar/anthropic/): The AI company behind Claude, built around making models safe and useful.
- [API](https://tidbit.legios.com.ar/api/): The menu of things one program lets another program ask for.
- [Token](https://tidbit.legios.com.ar/token/): The small chunk of text a language model actually reads.
- [Cloud computing](https://tidbit.legios.com.ar/cloud-computing/): Renting computers over the internet instead of buying your own.
- [Amazon Bedrock](https://tidbit.legios.com.ar/amazon-bedrock/): A single AWS service to call AI models from Anthropic, Meta, Amazon and others.
- [Google Cloud](https://tidbit.legios.com.ar/google-cloud/): Google's cloud business: computing, storage, data and AI services rented by the hour.
- [Artificial intelligence](https://tidbit.legios.com.ar/artificial-intelligence/): Getting computers to do things that, until recently, only people could do.
- [LLM](https://tidbit.legios.com.ar/llm/): A neural network trained on huge amounts of text to predict the next word.