You’ve heard of ChatGPT — but what is Claude? Claude is a whole family of AI models built by Anthropic, and it’s much more than just a chatbot. In this simple guide I’ll explain every Claude model — Haiku, Sonnet, Opus and Fable — and its products, Cowork and Code, with easy real-life examples. No jargon, for complete beginners.

What is Claude?

Claude is an AI assistant made by a company called Anthropic. Interestingly, the people who built Claude earlier worked at OpenAI (the makers of ChatGPT), and then started their own company to build their own AI.

But here’s the key thing — Claude isn’t just one tool. It’s a complete family: several AI models (something small, fast, balanced and powerful) plus products like Claude chat, Claude Cowork and Claude Code. Let’s meet the whole family, one by one.

The Claude family — picture four children in one home

Here’s a simple way to remember it. Imagine a family with four children, each good at different things. For small chores you don’t need the eldest; for a hard task you don’t hand it to the youngest. Claude works exactly like this — it has four public models, each suited to a different kind of job. Meet them:

Claude Haiku — the small, fast one

Haiku is the youngest child of the family. If you only need a small, quick job done, you don’t call the whole family — you send Haiku. It is the fastest and cheapest model, perfect for small, everyday tasks.

Claude Sonnet — the balanced all-rounder

Now imagine you’re going to the market to bring back some goods, but you don’t yet know whether the load will be heavy or light. You’d pick a balanced helper — one who can lift something a bit heavy, and also handle the light stuff easily. That’s Claude Sonnet.

Use Sonnet for writing emails, creating content, making documents, or everyday coding. Whenever you want a balanced model, it’s Sonnet — it’s fast, affordable and smart, all at once.

Claude Opus — the deep thinker

Think of your monthly household accounts when they get really tangled — how much came in, how much went out, where it went, and what’s left. This needs real thinking. Hand it to a small child and they’ll get confused; hand it to a balanced one and they still won’t do it well — because it needs deep thought.

This is where Claude’s most powerful model comes in — Opus. Use Opus only when your task is deep, difficult or needs a lot of reasoning — like solving a tangled problem, reading and analysing a big report, extremely tricky coding where one mistake ruins everything, or a big decision where many things must be weighed at once. Opus runs on understanding, not haste.

Mythos & Fable — the most powerful (and the restricted one)

In April 2026, India’s Finance Minister Nirmala Sitharaman called an emergency meeting with senior RBI officials and some big banks — and the topic was a new AI model called Mythos, built by Anthropic. Anthropic revealed that Mythos can find thousands of bugs in even the most powerful, sensitive or protected software — far better than humans — including hidden flaws the software makers themselves don’t know about, called zero-day vulnerabilities. The meeting’s worry was simple: as such models arrive, our systems must be strong enough that no one can launch a cyber attack.

Seeing this, Anthropic made a special decision: it did not release Mythos to the public — it restricted it to a few companies that can use it only after government approval. For the public, it released a model called Fable.

Think of Fable as the “bulldozer.” You don’t need a bulldozer to hammer a nail at home — a small hammer does it. But to build a highway or a bridge, you need heavy machinery. Fable is Mythos with safety guardrails added (so it can’t be misused), but its power is still enormous. Use it for extremely complex code, or analysing documents of 1,000–2,000 pages. That’s why scientists, experts and big companies use it. It’s very advanced — and very expensive, because it consumes a lot of tokens. So don’t use Fable just to write an email or read a short document.

And with that, we’ve met Claude’s four models — Haiku, Sonnet, Opus and Fable — plus the restricted sibling, Mythos.

Which Claude model should you use?

ModelBest forSpeed & cost
HaikuSmall, quick, everyday tasksFastest, cheapest
SonnetEmails, content, documents, daily codingBalanced — fast, affordable, smart
OpusDeep, complex thinking; hard problems; big analysisPowerful, but slower & pricier
FableExtremely complex code, huge documents (expert-level)Most powerful, most expensive
MythosSecurity/enterprise use — restrictedGovernment-approved companies only

Simple rule: for most everyday work, Sonnet. Quick small jobs, Haiku. Deep, hard thinking, Opus. Only reach for Fable when the task is genuinely heavy.

Beyond chat — meet Claude Cowork

So far we’ve talked about Claude that answers your questions — that’s Claude chat. But is that all AI can do? No. Say you want construction done at home. You grab a contractor and ask questions — how will this work, what’s the budget, what materials are needed. Asking those questions is like using Claude chat: ask, and get answers.

But the moment you hand over the job, Claude Cowork begins. Cowork doesn’t just chat — it takes actions. Like a contractor: if it’s a small job, it does it itself; if it’s big, it decides how many workers to bring and when to start, does the whole job, and at the end even makes sure the work you asked for was done properly.

For example: ask Claude chat “give me 5 business ideas,” pick one, and then tell Claude Cowork “make a full plan for this.” It will build you a complete plan — start to finish.

Claude Code — build without coding

People from the 90s will remember: to make a phone call, we’d go to an STD or PCO booth. Then the mobile phone arrived and changed the whole game. The same thing is now happening with coding.

Whenever you need a website, app or tool, what do you do? You go to a developer — who gives you a long timeline, then a long bill, then charges separately for every change. And if you don’t have technical knowledge, you can imagine how it goes. Claude Code changes all of this. On Claude Code, just like chatting on WhatsApp, you can create any website, app or tool without any technical knowledge — at a much lower cost than a developer, in far less time, with as many changes as you like.

And the live proof is me: I built my own product, brAInam.ai, using Claude Code — without any technical knowledge.

Why does this matter for you?

Because Claude isn’t just “another ChatGPT.” It’s a whole family — a model for every kind of task, and products (Cowork and Code) that let even a non-technical person get real work done and even build software. Once you know which Claude to use when, you work smarter, faster and cheaper. (Curious why different AIs behave differently? See why AI models behave differently.)

Frequently asked questions

What is Claude?

Claude is a family of AI models and products built by Anthropic — from a simple chat assistant to models that can write complex code and analyse huge documents.

Who made Claude?

Anthropic — a company started by people who earlier worked at OpenAI (the makers of ChatGPT).

What are the Claude models?

Four public models — Haiku (fast), Sonnet (balanced), Opus (powerful) and Fable (most powerful) — plus Mythos, a restricted model for approved organisations only.

Which Claude model is best for me?

For most everyday work, Sonnet. For quick, simple tasks, Haiku. For deep, complex thinking, Opus. Fable is only for expert-level heavy work.

What is Claude Cowork?

A version of Claude that doesn’t just chat but takes actions and completes whole tasks for you — like handing a job to a contractor.

What is Claude Code?

A tool to build websites, apps and tools just by describing what you want — no coding knowledge needed.