Claude

basicwidespread

Claude is the family of artificial intelligence models and conversational assistant from the US company Anthropic, which understands and writes text, answers questions, codes, and performs tasks independently.

In plain terms

Claude is an AI assistant built on a large language model: you enter your question or task in everyday language, including Hungarian, and it responds in text. It is like a well-read, patient colleague who is always available, but sometimes makes mistakes: it writes a letter, summarizes a long document, explains a contract, or helps consider a decision.

The name Claude refers to several things at once. It denotes the model family (the 'engine'), the chat application accessible in browsers, on computers, and on phones, and the products built on this engine, such as Claude Code for programmers. Several sizes of models exist: smaller ones are faster and cheaper, while larger ones are slower and more expensive but can handle more difficult tasks.

It is not the underlying operating principle that distinguishes it from competitors like ChatGPT or Gemini, but the emphasis. Anthropic prioritizes safety and predictable behavior, and Claude is mainly used for processing long texts, programming, and multi-step work performed as an agent.

How it works

The model was pre-trained on a huge amount of text, so it learned the patterns of the language and absorbed much knowledge about the world.
When you write to it, it breaks your text into small pieces (tokens), then estimates the likely continuation piece by piece based on the entire conversation. This is how the response is assembled.
During training, Anthropic also teaches the model based on a written set of principles on how to behave: be helpful and honest, and do not cooperate in harmful activities.
If it also receives tools (web search, file reading, program execution), it works as an agent: it plans the steps, executes them, checks the result, and corrects if necessary.
Companies can integrate it into their own systems via a programming interface (API), so you often use it without even seeing its name.

An everyday example

You receive a twenty-page insurance contract and do not understand the fine print. You upload it to Claude, ask when the insurer does not pay, and it summarizes the answer in bullet points in plain language, showing exactly which paragraph contains the information.

Who is affected

employeeIt can be used as a daily work tool for creating emails, summaries, tables, and presentations, transforming the workflow of office work. → Work & HR

developerClaude is the foundation of one of the most widespread coding agents, which independently writes, tests, and fixes program code. → Code

entrepreneur, executiveIt can be integrated into customer service, internal knowledge bases, or administrative processes, but you must decide on costs, data handling, and liability.

parent, student, teacherIn learning, it is useful as an explanatory and practice assistant, but it is also suitable for replacing homework, so it is worth setting rules for its use. → Education & Family

A common misconception

MythClaude is an omniscient search engine that always returns precise, up-to-date facts.

In realityClaude does not search a database but formulates answers based on learned patterns, so it can be convincingly wrong. Its knowledge is current only up to a specific date; it sees fresher information only if given web search access or a document.

Limits and risks

It may confidently state falsehoods (called hallucinations), so always verify important facts, numbers, and references.
It is a good starting point for medical, legal, and financial questions, but does not replace a specialist, and the responsibility for the decision remains with you.
Consider what personal or corporate data you provide to it, and check the data processing settings in your subscription.
If an agent acts independently (modifying files, sending emails, making purchases), its errors have real consequences, so human approval should remain for sensitive steps.

Similar, but not the same

Anthropic – Anthropic is the developer company, while Claude is the product, just as OpenAI is the company and ChatGPT is the product.
ChatGPT – ChatGPT is OpenAI's assistant with similar goals; a different company, different models, different strengths and rules.
Claude Code – Claude Code is a coding agent built on Claude models for programming, meaning it is an Anthropic product, not the model itself.

Related concepts

Broader

Large language model, Generative AI, MI-asszisztens, Foundation model

Narrower

Claude Code, Claude Opus, Claude Sonnet, Claude Haiku

On the edge

Anthropic, ChatGPT, Gemini, AI-agent, Model Context Protocol

A brief history

Anthropic was founded in 2021, and the first public version of Claude was released in March 2023. The Claude 3 family consisting of three sizes (Haiku, Sonnet, Opus) was introduced in 2024, and Claude Code became generally available in 2025.

Where things stand

At the end of 2026, Claude is one of the best-known AI assistants, used worldwide by individuals and companies alike. It is particularly strong in programming and in longer, independent work carried out as an agent.

Status: 2026 November

In the news

Appears in 207 Agent Age stories; the curve shows weekly occurrence in the news magazine.

Ethan Mollick publishes Portals, a browser-based 3D experience of 471 generated worlds built by agent teamsNov 9, 2026 · Code
Claude Code's 'Don't Ask Again' file exposes passwords and keys in 1,122 public copiesNov 7, 2026 · Code
Claude Code leaked users' email addresses in web search headers; Anthropic deployed a privacy fixNov 6, 2026 · Jog
Apple, Microsoft and Meta tighten internal AI usage limitsNov 5, 2026 · Code
Claude writes 80% of Anthropic's code; METR publishes independent assessment of its AI R&D accelerationNov 5, 2026 · Code
Postman added itself to Claude Code and Codex without asking permissionNov 5, 2026 · Code
Bug report alleges Claude Code writes user's email address into User-Agent headerNov 5, 2026 · Code
Visa test: AI shopping agents failed 42.8% of runsNov 5, 2026 · Money

Also known as

Claude · Claude AI · Anthropic Claude · Claude chatbot · Claude asszisztens · Claude modell · Claude AI agent · claude.ai

Sources

This entry was machine-edited and has undergone machine expert review; the “In the news” section updates automatically from Agent Age stories.