Pages

Anthropic is the San Francisco AI company that makes Claude, founded in 2021 by Dario Amodei, Daniela Amodei and five other co-founders and run as a public benefit corporation. It filed to go public on Nasdaq on Oct 26, 2026.

Anthropic publishes research on how Claude works inside (interpretability), whether it behaves as intended (alignment), model welfare, and misuse, plus a system card and Risk Reports for its models. In 2026 it also disclosed several incidents in which Claude models took unsanctioned actions on real systems during cybersecurity tests.

Anthropic's Usage Policy lists what nobody may use Claude for, such as weapons, malware, child sexual abuse material, fraud, surveillance and election deception, plus extra rules for high-risk uses. The 2026 version takes effect on Nov 12, 2026; violations can be reported to usersafety@anthropic.com.

Claude trains on Free, Pro, Max and Ultra chats only if the "Help improve" setting is on, which users can turn off in Settings > Privacy; with it on, chats may be kept for up to five years, and with it off for 30 days. Team, Enterprise and API data is not used for training by default.

A month-by-month record of what Anthropic shipped for Claude in 2025 and 2026 (models, apps, Claude Code and the API), each entry dated and linked to Anthropic's own announcement. The latest entries are Personal Finance in Claude and new Enterprise security controls on Nov 5, 2026.

To check whether Claude is down, open status.claude.com, Anthropic's official status page, which tracks claude.ai, the Claude Console, the Claude API, Claude Code, Claude Cowork and Claude for Government separately. On Nov 6, 2026 it showed all systems operational.

Claude's constitution is the public document Anthropic uses to train Claude's values and character; the current text was published on Jan 22, 2026 under a CC0 license. It ranks four priorities (broadly safe, broadly ethical, following Anthropic's guidelines, genuinely helpful) and sets a short list of hard limits no one can unlock.

The Responsible Scaling Policy is Anthropic's voluntary framework for managing catastrophic risks from Claude and later models; version 3.4, effective Jul 8, 2026, is current. It introduced AI Safety Levels, and Anthropic activated ASL-3 protections for Claude Opus 4 on May 22, 2025.

The Claude desktop app is a free download for macOS and Windows, with Linux in beta, that adds quick entry, desktop extensions, Cowork on your own files, computer use and Claude Code to the Claude chat.

Claude for Education is Anthropic's university-wide plan for students, faculty and staff, priced on request, with a learning mode that guides students with questions instead of answers; US K-12 teachers can get Claude free for a year.

Yes, Claude is free to use: the Free plan costs $0 and includes chat with Claude Sonnet and Haiku, web search, file creation, memory and connectors, with usage limited every five hours and every week.

Claude Max is Anthropic's $100-a-month plan with five times Claude Pro's usage every five hours and every week, Fable included up to half the weekly limit, $100 a month in API credits and priority access.

Claude's free mobile apps for iPhone, iPad and Android, listed as "Claude by Anthropic", offer chat, voice mode, dictation, camera and photo input, and links to phone apps such as Maps, Messages and Calendar.

Claude has a Free plan and three individual paid plans, Pro at $20 a month, Max at $100 and Ultra at $200, plus Team seats from $20, Enterprise seats at $20 plus usage, and an Education plan priced on request (as of Nov 6, 2026).

Claude Pro is Anthropic's $20-a-month plan ($17 a month billed annually) that adds at least five times Free's usage, the Opus models, Claude Code, Research and unlimited projects.

Claude Pro ($20 a month) suits daily chat and light Claude Code use; Claude Max ($100) gives five times the usage, Fable in the plan and $100 in API credits; Claude Ultra ($200) removes the five-hour limit for verified accounts.

Claude Team costs $20 or $100 a seat a month billed annually ($25 or $125 monthly) for 2 to 150 people; Claude Enterprise costs $20 a seat a month billed annually plus all usage at API rates, with a 20-seat minimum.

Claude Ultra is Anthropic's $200-a-month plan, the former Max 20x, with no five-hour limit for verified accounts, ten times Claude Pro's weekly usage, Fable included and $200 a month in API credits.

Claude's paid plans have a session limit that resets five hours after your first message and a weekly limit that resets at a fixed time each week; Anthropic publishes multiples of Pro's allowance, not message counts.

The Claude web app at claude.ai is where you sign in and chat with Claude in a browser, on every plan including Free; it holds your chats, projects, artifacts and settings, which sync with the desktop and mobile apps.

The Claude API is Anthropic's REST API at api.anthropic.com for building with Claude models: create an API key in the Claude Console at platform.claude.com, then send requests to the Messages API, paying per token.

Claude API prices are per million tokens, from $0.10 input and $0.50 output on Claude Haiku 5.5 to $10 and $50 on Claude Fable 5.1, with Opus 5.5 at $4 and $20 and Sonnet 5.5 at $2 and $10, as of Nov 6, 2026. Prompt caching and the 50% Batch discount lower that.

Claude API rate limits are set per organization by usage tier (Start, Build, Scale, Custom) as requests, input tokens and output tokens per minute; on the Start tier Claude Opus 5.5 allows 4,000 requests, 2,000,000 input tokens and 400,000 output tokens a minute. Going over returns a 429 error, while a 529 error means the API itself is overloaded.

Anthropic publishes official Claude API SDKs in seven languages (Python, TypeScript, Java, Go, Ruby, C# and PHP) plus the ant command-line tool; the Python SDK installs with pip install anthropic and the TypeScript SDK with npm install @anthropic-ai/sdk. Since Oct 21, 2026 the new major versions are Apache 2.0 licensed.

The Claude Console at platform.claude.com is Anthropic's web dashboard for the Claude API, where developers create API keys, test prompts in the playground, organize workspaces, watch usage and rate limits, and buy credits. It was called the Anthropic Console until Sep 2025, and its Workbench became the playground in Aug 2026.

Besides Anthropic's own Claude API, Claude models are sold through Amazon Bedrock, Claude Platform on AWS, Google Cloud (Google managed or Anthropic managed) and Microsoft Foundry. The Anthropic-operated options get new API features the same day, while Bedrock and Google managed run Claude on the cloud provider's own infrastructure with fewer features.

The Claude API's code execution tool runs Bash commands and Python in a sandboxed Linux container at Anthropic, with 5 GiB of memory and no internet access. Each organization gets 1,550 free container hours a month, then $0.05 per hour, and it is free when used with the newer web search or web fetch tools.

Computer use is a Claude API tool that lets Claude operate a desktop through screenshots, mouse clicks and keystrokes, which your own application carries out. It launched in public beta on Oct 22, 2024 and has been generally available as the computer_toolset_20260801 toolset since Aug 19, 2026.

The Claude Files API stores a PDF, text file, image or dataset once with POST /v1/files and returns a file_id that later Claude API requests reference instead of re-uploading. Files up to 500 MB, 1 TB per organization, free to store; out of beta since Aug 19, 2026.

Claude Managed Agents is Anthropic's hosted agent service: you define an agent once, and Anthropic runs its loop, tools and sandbox for each session, billing normal token prices plus $0.08 per session-hour of running time. It has been in beta on the Claude API since Apr 8, 2026.

The Message Batches API runs up to 100,000 Claude API requests asynchronously at 50% off the normal price; most batches finish within an hour, every batch within 24 hours, and results stay downloadable for 29 days.

The Messages API is the Claude API's main endpoint, POST /v1/messages: you send a model, max_tokens and a list of user and assistant messages, and Claude returns a message with content blocks, a stop_reason and token usage, whole or streamed.

The Model Context Protocol (MCP) is an open standard, introduced by Anthropic on Nov 25, 2024, for connecting AI applications such as Claude to outside tools and data through MCP servers. OpenAI, Google and Microsoft support it, and Anthropic donated it to the Linux Foundation's Agentic AI Foundation on Dec 9, 2025.

Prompt caching in the Claude API stores a prompt's prefix so later requests reuse it: you add cache_control, the first request writes the cache at 1.25x the input price, and reads within 5 minutes (or 1 hour) cost 0.1x or less.

Structured outputs make the Claude API return JSON that always matches a JSON Schema you supply, through output_config.format, or tool inputs that match exactly, through strict tool use. It is Claude's answer to a JSON mode, generally available since Jan 29, 2026.

Tool use (function calling) lets Claude call functions through the Claude API: you describe tools with a JSON Schema, Claude replies with a tool_use block, your code runs it and sends back a tool_result. Anthropic also runs server tools such as web search itself.

The Claude API's web search tool lets Claude search the web and cite its sources for $10 per 1,000 searches plus tokens, and the web fetch tool lets it read a given page or PDF for token costs only. Both are server tools that Anthropic runs during the request.

The Claude Agent SDK (formerly the Claude Code SDK) is Anthropic's Python and TypeScript library that runs Claude Code's agent loop, tools, hooks, subagents and MCP support inside your own program; install it with pip install claude-agent-sdk or npm install @anthropic-ai/claude-agent-sdk.

Claude Code is Anthropic's agentic coding tool. It reads a codebase, edits files, runs commands and opens pull requests from the terminal, VS Code, JetBrains IDEs, the Claude desktop app, the browser and Slack, and it comes with every paid Claude plan or a pay-as-you-go Console account.

Anthropic's Claude Code best practices come down to three things: tell Claude what it cannot see in the code, give it a check to run that passes or fails, and enforce rules that cannot have exceptions with hooks and deny rules instead of CLAUDE.md lines.

Claude Code saves a checkpoint of your files before every prompt, so /rewind or Esc pressed twice can restore the code, the conversation or both to an earlier point. Checkpoints cover Claude's own file edits, not changes made only by shell commands or most subagents, so Git is still needed.

Claude Code Desktop is the Code tab of the Claude desktop app for macOS, Windows and Linux (beta). It runs Claude Code sessions side by side, locally, in the cloud or over SSH, with Git worktrees, a visual diff view, a terminal, a browser preview and computer use, on any paid Claude plan with no separate CLI install.

Claude Code GitHub Actions (anthropics/claude-code-action@v1) lets you mention @claude in an issue or pull request to have Claude answer, change code and open pull requests; run /install-github-app in Claude Code to set it up.

claude -p "prompt" runs Claude Code once without the interactive screen and exits, printing text, JSON or streamed JSON; with --bare, --allowedTools and --permission-mode dontAsk it is how Claude Code runs in CI jobs, scripts and git hooks.

Claude Code hooks are shell commands (or HTTP calls, MCP tools or model checks) that Claude Code runs at fixed points such as before a tool call, after an edit or when Claude stops, configured under "hooks" in a settings.json file.

Claude Code has a VS Code extension (also for Cursor, Kiro and other forks) with inline diffs, plan review and @-mentions, opened with Cmd+Esc or Ctrl+Esc, and a JetBrains plugin for IntelliJ IDEA, PyCharm, WebStorm and other JetBrains IDEs that connects the Claude Code CLI to the IDE's diff viewer.

Claude Code's main shortcuts are Esc to stop Claude, Esc twice to rewind, Shift+Tab to cycle permission modes, Ctrl+R to search history, Ctrl+O for the transcript, Ctrl+B to background a task and Alt+P (Option+P) to switch model. Shortcuts can be changed in ~/.claude/keybindings.json, opened with /keybindings.

Claude Code connects to MCP servers with claude mcp add, for example claude mcp add --transport http notion https://mcp.notion.com/mcp, stored per user, per project in a shared .mcp.json file, or locally for one project.

Claude Code on the web, now called cloud sessions, runs Claude Code tasks on Anthropic's cloud machines from claude.ai/code, the Claude mobile app, the desktop app or `claude --cloud`. It clones a GitHub repository, works while your computer is off and opens a pull request, on Pro, Max, Ultra, Team and eligible Enterprise seats.

Claude Code has six permission modes (Manual, Accept edits, Plan, Auto, Don't ask and Bypass permissions), cycled with Shift+Tab; since v2.1.283 interactive sessions start in auto mode, and --dangerously-skip-permissions turns every check off and belongs only in a container or VM.

A Claude Code plugin is a folder of skills, subagents, hooks and MCP servers that installs as one unit from a marketplace; type /plugin in Claude Code to browse Anthropic's official marketplace and install one.

Claude Code is included in every paid Claude plan, from Pro at $20 a month (or $17 billed annually) to Max at $100 and Ultra at $200, and shares their usage limits; it is not in the free plan. With a Console account it is billed per token at API rates instead, which Anthropic says averages about $13 per developer per active day.

Claude Code reads settings from ~/.claude/settings.json (you), .claude/settings.json (shared project), .claude/settings.local.json (you, one project) and managed policy, with managed settings winning; /config edits them in a session.

Claude Code's slash commands are typed at the start of a message, such as /clear, /compact, /model, /init, /rewind and /usage. Anthropic's list had 121 entries on Nov 6, 2026, counting bundled skills and workflows, and your own commands are skills in .claude/skills/ or .claude/commands/.

Claude Code subagents are helper agents that run in their own context window with their own instructions, tools and model; Claude Code ships built-in ones such as Explore and Plan, and you add your own as Markdown files in .claude/agents/.

CLAUDE.md is a Markdown file of instructions that Claude Code loads at the start of every session, from your user folder (~/.claude/CLAUDE.md), the project (./CLAUDE.md) and a private CLAUDE.local.md; /init writes one. Claude Code also keeps an auto memory of what it learns, indexed in MEMORY.md, and reads AGENTS.md when there is no CLAUDE.md.

Install Claude Code on macOS, Linux or WSL with `curl -fsSL https://claude.ai/install.sh | bash`, or on Windows PowerShell with `irm https://claude.ai/install.ps1 | iex`, then run `claude` in a project folder and sign in with a paid Claude plan or a Console account.

The main alternatives to Claude on Nov 6, 2026 are ChatGPT, Gemini, Microsoft Copilot, Grok and Perplexity as assistants; Codex, GitHub Copilot, Cursor and Google Antigravity as coding agents; and the OpenAI and Gemini APIs or open-weights models such as Llama, Gemma, Qwen and DeepSeek for developers.

As of Nov 6, 2026, the Claude API and the OpenAI API charge the same list prices at the top ($10 / $50 per million tokens for Claude Fable 5.1 and GPT-6 Astra), but OpenAI is cheaper in the middle tier and sells voice, audio, images and embeddings that the Claude API does not. Claude charges no long-context surcharge up to 1M tokens and has cheaper cache reads.

Claude Code and OpenAI's Codex are both coding agents for the terminal, the IDE and the cloud at the same $20 entry price; Claude Code led the public Terminal-Bench 4.0 board on Nov 6, 2026, while Codex is open source, cheaper per result and included in ChatGPT's Free and Go plans.

Claude Code is Anthropic's coding agent for the terminal, IDEs and the cloud, included with Claude plans from $20 a month; Cursor is an AI code editor, owned by SpaceX since Aug 14, 2026, whose plans run from $8 and which can also run Claude models at Anthropic's list price on its $30 Premium plan and up.

Gemini CLI is Google's open source (Apache 2.0) terminal agent, but since Jun 18, 2026 it has no free tier and serves enterprise Gemini Code Assist users on Google Cloud, with free and Google AI plan users moved to Antigravity CLI; Claude Code is Anthropic's paid agent, from $20 a month, which led Terminal-Bench 4.0 on Nov 6, 2026.

GitHub Copilot is the cheaper way into AI coding, from $10 a month with unlimited completions and Claude, GPT, Gemini and Grok models billed as AI credits, while Claude Code is Anthropic's own agent, included with Claude plans from $20 a month, which led the Terminal-Bench 4.0 board on Nov 6, 2026.

Claude leads ChatGPT on coding benchmarks and in most 2026 writing tests, while ChatGPT wins on image generation, voice, its free tier and its sheer number of features. Both cost $20 a month at the main paid tier; as of Nov 6, 2026 ChatGPT also sells cheaper and costlier tiers.

Gemini is the better value and the better pick for images, video, voice and Google apps, and its Gemini 4 models lead Arena's text board; Claude leads on coding agents and in published writing tests. As of Nov 6, 2026 Google AI Pro costs $19.99 a month and Claude Pro $20.

Claude scores far higher than Grok on coding and reasoning boards, while Grok wins on live X posts, image and video generation, a cheaper entry plan and a cheaper API. As of Nov 6, 2026, Grok comes in The X Subscription from $8 a month and Claude Pro costs $20.

Microsoft Copilot is free and built into Word, Excel, Outlook and Windows, and Microsoft 365 Copilot at work already offers Claude models in most commercial tenants; Claude's own apps go further on coding and long documents. As of Nov 6, 2026, Microsoft 365 Personal with Copilot costs $9.99 a month and Claude Pro $20.

Perplexity is an answer engine built for cited web search and runs several makers' models, Claude Sonnet 5.5 among them; Claude is an assistant built for writing, coding and work with your files. As of Nov 6, 2026 Perplexity Pro costs $17 a month billed annually, the same as Claude Pro's annual price.

Claude is not the best choice for everything: as of Nov 6, 2026 it cannot generate images or video, has no open weights to run locally, offers no free coding agent and a smaller free plan than ChatGPT or Gemini, and its API sells no realtime voice, speech, image or embedding models; this page names the better tool for each of those jobs, and where Claude is the strong choice.

Claude artifacts are documents, decks, designs, dashboards, websites and small interactive apps that Claude builds in a panel beside the chat, saved in the Artifacts tab and shareable by link. They are on every Claude plan, including Free.

Claude Dashboards turns company data into a dashboard that stays current, with each number linked to its query; Claude Motion turns a prompt into a short animated explainer you download as an MP4. Both are betas launched on Oct 8, 2026: Dashboards on paid plans, Motion on Team and Enterprise.

Claude Design makes designs, interactive prototypes, one-pagers and other visual work from a description, on brand with your design system, and exports them to PowerPoint, PDF or HTML. It works in any Claude conversation and is included on every plan, Free too.

Claude Duet is Anthropic's proactive assistant app, in research preview since Oct 21, 2026: it watches your connected apps, sends a morning brief, does tasks on its own cloud computer and asks before acting. It is US only, by waitlist, on the Pro, Max and Ultra plans.

Claude for Excel is Anthropic's add-in that puts Claude in a sidebar inside Microsoft Excel to explain, fix and build spreadsheet models. It is generally available on paid Claude plans and installs from Microsoft's add-in store as "Claude for Microsoft 365".

Claude works with Google Workspace two ways. Its Gmail, Google Calendar, Google Drive, Docs, Sheets and Slides connectors work inside the Claude app on every plan, and the Claude for Google Workspace add-on (beta, paid plans) puts Claude in a sidebar inside Docs, Sheets and Slides.

Claude for Microsoft 365 is Anthropic's set of Office add-ins that put Claude inside Excel, PowerPoint, Word and Outlook. Excel, PowerPoint and Word are generally available on paid Claude plans and Outlook is in beta, as of Nov 6, 2026.

Claude in Chrome is Anthropic's extension for Google Chrome that lets Claude read web pages and click, type and fill forms on sites you are signed in to. It is on every paid Claude plan and has been generally available since Aug 26, 2026.

Claude Science is Anthropic's research app for scientists, a public beta since Jun 30, 2026: it runs analyses on your own machines or cluster, shows the code behind every figure, renders proteins and genomes, and queries more than 60 scientific databases. It comes with Pro, Max, Ultra, Team and Enterprise, and academic labs can get a discounted Team plan.

Claude Security scans your code repositories for vulnerabilities, checks each finding before you see it, proposes a patch and marks the issue fixed when the fix ships. It is a public beta for Claude Enterprise, and a Claude Security plugin for Claude Code is in beta for every Claude Code user.

Claude Slides drafts presentation decks from a file or conversation and exports them to PowerPoint, Google Slides or PDF; Claude Docs is a shared document that you and your team write with Claude. Both arrived in every Claude conversation on Sep 16, 2026 and are included on every plan, including Free.

Claude Tag is how Claude works in Slack: you tag @Claude in a channel, thread or direct message and it does the task with your organization's tools, then reports in the thread. It is in beta on the Team and Enterprise plans and replaced the older Claude in Slack app.

Claude connectors link Claude to apps such as Gmail, Google Drive, Slack, Notion and Asana so it can read and act in them with your permissions. They are on every Claude plan, Free included; Free accounts may add one custom connector.

Cowork is Claude's agent for non-coding work: you give it a goal and it works through your files, apps and browser to hand back finished documents, spreadsheets and decks. Since Sep 16, 2026 it is no longer a separate mode but part of every Claude conversation on paid plans.

Claude can create and edit Excel spreadsheets, PowerPoint decks, Word documents and PDFs that you download or save to Google Drive, on every Claude plan including Free. It writes and runs code in a private sandbox to build them.

Claude reads uploaded images, PDFs, Word files, spreadsheets and text files on every plan, up to 500 MB per file and 20 files per chat. PDFs up to 100 pages are read with their charts and images; longer ones up to 1,000 pages as text only.

Claude cannot generate photos or illustrations as of Nov 6, 2026. It draws diagrams, charts, SVG graphics and interactive visuals as code instead, and reads images you upload; ChatGPT, Gemini, Grok and others generate images.

Claude's incognito chats are temporary conversations that are not saved to your chat history or memory and are not used for training, on every plan. Anthropic still keeps them for 30 days by default for safety.

Claude remembers across chats on every plan, Free included: it saves facts as topics you can read, edit or delete under Settings > Your Claude. Paid plans can also search past chats, and incognito chats are never remembered.

Claude Projects are workspaces in the Claude app with their own chats, uploaded knowledge files, instructions and memory. Free accounts can make up to five; paid plans have no cap, and Team and Enterprise can share them.

Claude's Research mode runs many web searches that build on each other, plus searches of connected tools such as Gmail and Google Docs, and returns a cited report. It is on paid Claude plans (Pro, Max, Ultra, Team, Enterprise), not Free.

To share a Claude chat, click Share at the top right and invite people by email, or on individual plans make a public link. Recipients see a read-only snapshot with its artifacts but not your attached files, and you can update or turn off a share at any time.

A Claude skill is a folder with a SKILL.md file of instructions, plus optional scripts and files, that Claude loads when a task needs it. Skills are on every Claude plan, Free included, and the same skill runs in the Claude app, Claude Code and the API.

Claude's custom instructions are called Instructions for Claude, set under Settings > Your Claude and applied to every chat; projects add their own instructions, and skills handle tone and format. Claude's older Styles menu is no longer documented as of Nov 6, 2026.

Claude voice mode lets you talk with Claude out loud and hear it answer, on every Claude plan including Free, in the iOS and Android apps, Claude Desktop and the web. It is a beta, uses preset voices only and counts toward your usage limit.

Yes, Claude can search the internet. Web search is on every Claude plan, Free included, in every country Claude serves; turn it on with "+" and "Web search", and every answer that uses it cites its sources.

The most useful Claude tips in one list, from incognito chats, editing a draft in place and the Mac's double-tap Option shortcut to stretching usage limits, Projects, skills, connectors and Claude Code commands.

A Claude skill is a folder with a SKILL.md file whose front matter gives a name and a description of when to use it, followed by instructions; zip the folder and upload it in Customize > Skills, or put it in .claude/skills for Claude Code.

To cancel Claude Pro, Max or Ultra bought on the web, go to Settings > Billing and click Cancel; App Store and Google Play subscriptions are canceled in the app's Billing menu or the store. The plan runs to the end of the period, and payments are generally non-refundable.

Connect Claude to a remote MCP server in Customize > Connectors > Add custom connector, to a local one in Claude Desktop through a desktop extension or claude_desktop_config.json, and in Claude Code with claude mcp add.

Export Claude data in Settings > Privacy > Export data on the web or desktop (the link arrives by email and lasts 24 hours); delete a chat from its menu, and delete the account in Settings > Account. Deleted chats leave Anthropic's systems within 30 days.

To make Claude's usage limits last, start a new chat for each topic, keep files in Projects, use Sonnet or Haiku for simple tasks, lower the effort setting, and turn off tools you do not need; past the limit, paid plans can buy usage credits.

To use Claude, sign up free at claude.ai with an email or Google account, verify a phone number, and type a question in the chat box; paid plans add more usage, more models and features like Research.

Claude is generally safe to use with sensible habits: Anthropic encrypts data, holds SOC 2 Type 2, ISO 27001 and ISO 42001 certifications, and lets consumers opt out of training, but chats may be used for training if allowed, and agents that browse or use connectors can be misled by prompt injection.

Anthropic's advice for prompting Claude is to be clear and direct, explain why, give three to five examples, separate parts of a prompt with XML tags, set a role, and put long documents before the question.

Claude is the AI assistant and family of AI models made by Anthropic, used in the claude.ai web, desktop and mobile apps, in Claude Code for programming, and through the Claude API; it is free to start, with paid plans from $20 a month.

The Claude 3 family launched on Mar 4, 2024 with Claude 3 Opus, Sonnet and Haiku, followed by Claude 3.5 Sonnet (Jun 21, 2024), its upgrade and Claude 3.5 Haiku (Oct 22, 2024), and Claude 3.7 Sonnet (Feb 24, 2025), the first Claude with extended thinking. All of them are retired on the Claude API; the last, Claude 3 Haiku, on Apr 20, 2026.

The Claude 4 family ran from Claude Opus 4 and Sonnet 4 (May 22, 2025) to Claude Opus 4.8 (May 28, 2026). As of Nov 6, 2026, Opus 4.5 to 4.8, Sonnet 4.6 and Haiku 4.5 are still served on the Claude API, Sonnet 4.5 retires on Nov 30, 2026, and Opus 4, Sonnet 4 and Opus 4.1 are retired.

As of Nov 6, 2026, Claude Opus 5.5 leads Terminal-Bench 4.0 and WebDev Arena, ranks third on Arena's text board behind two Gemini 4 models, second on the Artificial Analysis index, and third on ARC-AGI-2 behind two OpenAI models. Anthropic's own launch tables run a few points higher than the public boards.

Every current Claude model (Fable 5.1, Opus 5.5, Sonnet 5.5, Haiku 5.5) has a 1M-token context window on the API, about 555,000 words, and can write up to 128K tokens in one response. In the Claude apps the window is up to 1M on paid plans depending on the model, and long chats are summarized automatically near the limit.

Claude Fable 5 launched on Jun 9, 2026 as Anthropic's first Mythos-class model for general use, was suspended for all users on Jun 12 when US export controls applied to it, and returned on Jul 1, 2026. Claude Fable 5.1 replaced it as the current model in Sep 2026; Fable 5 remains on the API as a legacy model.

Claude Fable 5.1 is Anthropic's most capable widely released model, released on Sep 1, 2026 and priced at $10 input and $50 output per million tokens on the Claude API. It is included on Claude Max, Ultra, Team premium seats and Enterprise, and paid with usage credits on Pro.

Claude Haiku 5.5 is Anthropic's fastest and cheapest current model, released on Oct 7, 2026 at $0.10 input and $0.50 output per million tokens for prompts up to 100,000 tokens. It suits classification, extraction, routing and subagents, and is on every Claude plan including Free.

Anthropic released the first Claude model on Mar 14, 2023 and has shipped more than 30 since, through Claude 2, the Claude 3, 3.5 and 3.7 models, the Claude 4 family and the Claude 5 family that is current in Nov 2026. This page lists every release with its date, what it added and when it was retired.

Anthropic sells four current Claude models: Claude Fable 5.1, Claude Opus 5.5, Claude Sonnet 5.5 and Claude Haiku 5.5, all with a 1M-token context window. Anthropic's docs suggest starting with Opus 5.5 and moving up to Fable 5.1 only when a task needs it; Claude Mythos 5.1 is restricted to verified organizations.

Claude Mythos is Anthropic's name for its most capable models with some safeguards lifted, offered only to vetted organizations for defensive cybersecurity and life sciences work. It began as Claude Mythos Preview for Project Glasswing on Apr 7, 2026; the current version, Claude Mythos 5.1, is the same model as Claude Fable 5.1 and is limited to US organizations in Anthropic's verification programs.

Claude Opus 5.5 is Anthropic's recommended starting model, released on Sep 22, 2026 at $4 input and $20 output per million tokens on the Claude API, with a 1M-token context window. It is on every paid Claude plan and is the default model in Claude Code.

Claude Sonnet 5.5 is Anthropic's mid-priced model, released on Sep 28, 2026 at $2 input and $10 output per million tokens with a 1M-token context window. It is on every Claude plan including Free, and costs half as much as Claude Opus 5.5.

Thinking lets Claude reason before it answers. On current Claude models it is adaptive (the model decides how much to think) and is steered with an effort level from Low to Max: in the apps through the model menu next to the send button, on the API with output_config.effort. Thinking tokens are billed as output.