Claude by Anthropic: the complete guide
Claude is the artificial-intelligence assistant built by Anthropic. This guide covers the essentials to understand Claude — its models, Claude Code, the API, its approach to safety — and how to follow its news day to day. For the latest announcements, browse the Claude news feed.
The Claude model family: Opus, Sonnet, Haiku
Claude comes in three complementary tiers. Opus is the most capable model, built for complex reasoning, deep analysis and hard coding. Sonnet offers the best balance of performance, speed and cost — the everyday model. Haiku is the fastest and most affordable, ideal for high-volume and real-time applications. Each generation carries a version number; follow the Models category for the latest releases.
Claude Code: the developer agent
Claude Code is Anthropic's coding agent. It runs in the terminal, in IDEs, in the browser and as an app: it reads your codebase, runs commands, edits files, runs tests and automates entire engineering tasks — always under your control. Find its updates in the Code category.
The Claude API, SDKs and the MCP protocol
Developers integrate Claude via the Anthropic API and its SDKs (Python, TypeScript). You can force structured outputs, wire up tools (function calling) and connect data sources through MCP (Model Context Protocol), now an interoperability standard. Claude News itself exposes a search engine and an MCP server you can connect to your Claude.
Safety and Constitutional AI
Anthropic is first and foremost an AI safety company. Its Constitutional AI approach trains Claude to follow a set of principles for helpful, honest and harmless responses. Research and safety topics are covered in the Security and Research categories.
Claude vs ChatGPT and Gemini
Claude (Anthropic), ChatGPT (OpenAI) and Gemini (Google) are the three leading AI assistants. Claude stands out for writing quality, long-document analysis, reasoning rigor and agentic coding via Claude Code. The right choice depends on your use case: test each on the tasks that matter to you.
Learn: the Claude certification
To level up, Claude News offers a free Academy that prepares you for Anthropic's official certification, with domain modules, graded quizzes and a timed mock exam.
How to follow Claude news
Claude News continuously aggregates the press, Hacker News, Reddit and Anthropic's GitHub releases, then Claude summarizes and classifies each item. Browse the real-time feed, filter by category, subscribe to the RSS feed, or connect our MCP server to your own Claude to query the news straight from your conversations.
Go further
Frequently asked questions
What is Claude?
Claude is the AI assistant built by Anthropic. It is a large language model that can chat, write, analyze documents, write and debug code, reason through complex problems and use tools. Claude is available through the claude.ai web app, desktop and mobile apps, a developer API, and the Claude Code command-line tool.
Who is Anthropic?
Anthropic is the AI safety and research company that builds Claude. Founded by former OpenAI leaders, it focuses on building reliable, interpretable and steerable AI systems, notably through its 'Constitutional AI' approach.
What are the Claude models (Opus, Sonnet, Haiku)?
The Claude family comes in three tiers: Opus (the most capable, for complex reasoning and hard coding), Sonnet (the balance of performance, speed and cost — the everyday workhorse) and Haiku (the fastest and most affordable, for high-volume tasks). Each generation carries a version number; for the latest models and announcements, check the Claude News feed.
What is Claude Code?
Claude Code is Anthropic's coding agent that lives in your terminal (and in IDEs, the browser and the app). It reads your codebase, runs commands, edits files, runs tests and automates engineering tasks end to end, under your control.
How do I use the Claude API?
Developers integrate Claude via Anthropic's API (with Python/TypeScript SDKs). You send messages to the model, can force structured outputs, wire up tools (function calling) and connect data sources via the MCP (Model Context Protocol). Claude News has a dedicated category for product, API and SDK news.
Is Claude free?
Claude has a free tier on claude.ai with usage limits, plus paid plans (Pro, Max, Team, Enterprise) that raise quotas and unlock features. The API is billed per usage (per token).
Claude vs ChatGPT: which should I choose?
Claude (Anthropic) and ChatGPT (OpenAI) are the two leading assistants, with Gemini (Google) rounding out the trio. Claude is especially valued for high-quality writing, long-document analysis, reasoning and agentic coding via Claude Code. The best choice depends on your use case — compare them on the tasks that matter to you.
What is Constitutional AI?
It is Anthropic's method for aligning Claude: rather than relying only on human feedback, the model is trained to follow a set of principles (a 'constitution') that steer helpful, honest and harmless responses.
How can I follow Claude news in real time?
Claude News continuously aggregates the press, Hacker News, Reddit and Anthropic's GitHub releases, then Claude summarizes and classifies the information. Browse the categories (Models, Product, Code, Security, Research, Business), subscribe to the RSS feed, or connect our MCP server to your own Claude.
Claude News is an independent publication, not affiliated with Anthropic. 'Claude' and 'Anthropic' are trademarks of their respective owners.