C

Claude

Visit Claude

Claude is Anthropic's AI, built for problem solvers. Tackle complex challenges, analyze data, write code, and think through your hardest work.

No ratings yet|3

From $0

BusinessCodeCreativeGeneralKnowledgeResearch
Claude 首屏预览
On this page

What is Claude?

Claude is Anthropic's family of AI models and the Claude.ai app built around them, used for conversation, writing, coding, research, and increasingly autonomous agent work, across web, iOS, Android, desktop, and a growing set of specialized surfaces. Anthropic was founded in January 2021 in San Francisco by a group of former OpenAI researchers, including siblings Daniela Amodei (President) and Dario Amodei (CEO), with AI safety as a founding focus. In May 2026 the company closed a $65 billion Series H round at a $965 billion post-money valuation, and confidentially filed for an IPO the following month.

The current model lineup is the Claude 5 family: Claude Opus 5 (released July 24, 2026, built for complex agentic coding and enterprise work, with a 1M-token context window and adaptive thinking by default), Claude Sonnet 5 (the practical default for many agent and production workflows), Claude Fable 5 (a "Mythos-class" model released June 9, 2026 for long-running agents, difficult coding, and complex research), and Claude Haiku 4.5 (the fast, low-cost model).

Beyond chat, Anthropic has expanded Claude into several dedicated products built on the same models: Claude Code (an agentic coding tool for the terminal, desktop, browser, and IDEs), Claude Cowork (a desktop agent workspace for knowledge workers, launched in research preview in January 2026), Claude Design (launched April 2026, for turning prompts into prototypes, decks, and marketing visuals), and Claude Science (launched June 2026, a research workbench connecting to 60+ scientific databases, alongside an internal drug-discovery program Anthropic announced at the same time). Claude also supports Projects, reusable Skills, Artifacts, and connectors via the Model Context Protocol (MCP), an open standard Anthropic introduced for linking models to external tools and data.

Core Features

Claude 5 Model Family

Opus 5, Sonnet 5, Fable 5, and Haiku 4.5 span maximum capability, everyday-agent performance, long-running/complex-research work, and fast low-cost use, with up to a 1M-token context window on the top models.

Claude Code

An agentic coding tool available in the terminal, desktop app, browser, and supported IDEs, built to read, write, and run code as part of a development workflow.

Claude Cowork

A desktop agent workspace for knowledge workers who aren't in a terminal or code editor, launched in research preview in January 2026.

Claude Design

Turns a plain-language description into prototypes, slide decks, one-pagers, or marketing visuals, and can apply a team's existing design system for consistency.

Claude Science

A research workbench connecting to 60+ scientific databases, runnable on local machines, over SSH, or on an HPC login node so sensitive data stays on a lab's own infrastructure.

Projects, Skills & Artifacts

Projects group related conversations and files with shared context; Skills add reusable, specific capabilities; Artifacts generate and display standalone content like code, documents, or interactive pages alongside the conversation.

Model Context Protocol (MCP)

An open standard Anthropic introduced for connecting Claude to external tools, data sources, and services, powering connectors like Slack, Google Workspace, and Microsoft 365.

How to Use Claude

  1. Sign up at claude.ai, or download the iOS, Android, or desktop app — the Free plan requires no payment and includes the full Claude 5 model family at reduced weekly usage limits.
  2. Start a conversation, upload a file or document, or attach content for Claude to work from.
  3. Use Projects to group related conversations and files together with shared instructions and context, and reusable Skills to give Claude specific, repeatable capabilities.
  4. For software development, use Claude Code from the terminal, an IDE, the desktop app, or the browser, or reach for Claude Cowork for an agentic workspace built around non-coding tasks.
  5. For visual work, use Claude Design to turn a description into a prototype, slide deck, one-pager, or marketing asset, optionally grounded in your team's existing design system.
  6. For scientific or data-heavy research, use Claude Science, which can run locally, over SSH, or on an HPC login node so sensitive datasets stay on your own infrastructure while connecting to 60+ scientific databases.
  7. Connect external tools and data sources via MCP connectors (Slack, Google Workspace, Microsoft 365, and others), and upgrade to Pro, Max, Team, or Enterprise for higher usage limits and additional admin controls as needs grow.

Use Cases

  • General-purpose writing, research, analysis, and conversation, with memory and web search available even on the Free plan.
  • Agentic software development via Claude Code, from a terminal, IDE, browser, or the desktop app, for tasks ranging from a single bug fix to multi-file feature work.
  • Non-coding knowledge work handled by an agent via Claude Cowork — research, document assembly, and multi-step tasks for people who don't live in a terminal or code editor.
  • Turning a plain-language description into a prototype, slide deck, or marketing visual with Claude Design, optionally matched to a company's existing design system.
  • Computational science and biopharma research via Claude Science, connecting to 60+ scientific databases while keeping large or sensitive datasets on a lab's own infrastructure.
  • Building custom autonomous agents on top of Claude using the Claude Agent SDK, the same underlying harness that powers Claude Code.
  • Team and enterprise deployments with centralized billing, SSO, audit logs, and compliance controls via Team and Enterprise plans.

Pros & Cons

Pros

  • Claude spans one of the widest surface areas among AI assistants — chat, coding (Claude Code), a non-coding agent workspace (Cowork), design (Claude Design), and scientific research (Claude Science) — all built on the same underlying model family rather than separate, disconnected products
  • The Free plan already includes the full Claude 5 model lineup (Opus, Sonnet, Fable, Haiku), memory, web search, and file creation, just at reduced weekly usage limits rather than being locked to a lower-tier model
  • Claude Science and Claude Cowork are both built to run against a lab's or company's own infrastructure (local machines, SSH, HPC login nodes) with no-training defaults, which matters for teams with sensitive data
  • Anthropic introduced and open-sourced the Model Context Protocol (MCP), which has been adopted well beyond Claude itself, giving Claude a broad and growing ecosystem of third-party tool integrations

Cons

  • The product lineup has expanded quickly — Cowork, Claude Design, and Claude Science all launched within the first seven months of 2026 — and several remain in beta or research preview, so features and limits are still actively changing
  • Pricing has a wide gap between Pro ($17–$20/month) and Max (from $100/month for 5x or 20x usage), with no smaller step in between for someone who outgrows Pro's weekly limits but doesn't need a full Max-tier commitment
  • Model naming has shifted across the 4.x and 5.x generations (Haiku 4.5 alongside Opus 5, Sonnet 5, and Fable 5), and terminology like "Mythos-class" for Fable isn't self-explanatory, which can make it harder for a new user to know which model fits their task
  • Anthropic's valuation and funding have grown extremely fast — a $965 billion post-money valuation as of its May 2026 Series H, ahead of a confidentially filed IPO — worth factoring in given how young and fast-changing the company's broader product surface still is

Pricing

Free

$0

  • Full Claude 5 model family (Opus, Sonnet, Fable, Haiku) at 50% of Pro's weekly usage limits
  • Web, iOS, Android, and desktop chat
  • Memory, web search, and file creation
  • 200K context window
  • No training on your content

Pro

$17/month billed annually, or $20/month billed monthly

  • Everything in Free, plus Claude Code, Cowork, Design, and Science
  • More usage than Free, unlimited Projects
  • Research access
  • Microsoft 365 integration

Max

From $100/month (5x or 20x usage tier)

  • Everything in Pro
  • 5x or 20x more usage than Pro
  • Higher output limits
  • Early access to advanced features
  • Priority access during high traffic

Team

Standard seat $20/month annual ($25 monthly); Premium seat $100/month annual ($125 monthly) — 2–150 people

  • All Claude features with more usage than Pro
  • Central billing, SSO, admin controls
  • Role-based access, no training by default
  • Mix and match seat types

Enterprise (Self-Serve)

$20/seat + usage at API rates

  • Org/user spend limits, fine-grained RBAC
  • SCIM, audit logs, Compliance API
  • Custom data retention, IP allowlisting
  • HIPAA-ready option

Enterprise (Sales-Assisted)

Custom pricing (MSA/PO options)

  • Tailored for large organizations
  • Usage commitments and product bundling

Frequently Asked Questions