Organization
Anthropic
Anthropic is an AI model developer and research lab headquartered in San Francisco, California53. This catalog lists 1 software project and 1 evaluation tool from Anthropic, under licenses including the MIT License and the Apache License 2.0.
Documented roles: Model developer · Research lab · Developer platform · Consumer products
Anthropic, which describes itself as an AI safety and research company, develops the Claude family of models. It offers Claude through web, desktop, and mobile apps, through the Claude Platform API, and through the Claude Code coding agent.1789
Status: On October 12, 2026 USASI added the Claude Cowork research-preview launch date, from Anthropic's January 13, 2026 post introducing Labs, to this profile.64
Anthropic in depth
Anthropic PBC describes itself as an AI safety and research company that builds AI systems people can rely on and studies the opportunities and risks of AI. It is a public benefit corporation whose stated purpose is "the responsible development and maintenance of advanced AI for the long-term benefit of humanity." Its privacy policy lists Anthropic PBC at a San Francisco, California address.13
Anthropic develops the Claude family of AI models and offers them as hosted services; its Hugging Face organization lists no model repositories. As of October 2026 its model documentation lists four current models (Claude Fable 5.1, Claude Opus 5.5, Claude Sonnet 5.5, and Claude Haiku 5.5), each with a 1M-token context window. A separate model, Claude Mythos 5.1, has the same capabilities as Claude Fable 5.1 but is offered only to organizations that pass Anthropic's verification programs.142117
People can use Claude through the Claude app on the web, desktop, and mobile, through the Claude Code coding agent, and through the Claude Platform API. Developers can also reach Claude models on Amazon Bedrock, Google Cloud, Microsoft Foundry, and Claude Platform on AWS. Subscription plans range from Free to Enterprise; this profile does not list prices.60737612
Anthropic publishes a Responsible Scaling Policy (version 3.4 took effect on July 8, 2026), system cards for its models, periodic risk reports, a Transparency Hub, a Usage Policy, and a public constitution that describes the values it intends Claude to hold. A Long-Term Benefit Trust, which holds a special class of stock, takes part in electing the company's board.4950521621
Anthropic has offered Claude to U.S. government agencies, including custom Claude Gov models for national security customers. In 2026 the Department of War excluded Claude from its supply chain after a contract dispute over use restrictions; court records show that a federal district court set aside one designation in August 2026, while the D.C. Circuit denied Anthropic's challenge to a separate exclusion on September 25, 2026.8085
What Anthropic does
Anthropic says it is dedicated to building AI systems that people can rely on and to producing research about the opportunities and risks of AI.1
- Its main product line is Claude, a family of large language models that take text and images as input and produce text. Anthropic offers the models as hosted services; its Hugging Face organization lists datasets but no model repositories.1417
- Anthropic sells Claude plans for individuals (Free, Pro, and Max), for organizations (Team and Enterprise), and for schools (Education), and it offers pay-as-you-go API access on the Claude Platform.12
- Its research page names five research teams: Alignment, Economics, Frontier Red Team (which studies what new models mean for cybersecurity, biosecurity, and autonomous systems), Interpretability (which studies how models work inside), and Societal Impacts.56
- Anthropic introduced the Model Context Protocol (MCP) in November 2024 as an open standard for connecting AI applications to outside data and tools, and in December 2025 donated it to the Agentic AI Foundation, a directed fund under the Linux Foundation.5859
- Since September 2026 Anthropic has run verification programs for cybersecurity and life sciences work. Vetted organizations in these programs get access to models with safeguards adjusted for their field, including Claude Mythos 5.1.8990
- Anthropic also works with governments, including Claude Gov models for U.S. national security customers and commitments announced in October 2026 to support the federal Genesis Mission for AI-driven science.8086
Claude models
Anthropic's models overview lists Claude Fable 5.1, Claude Opus 5.5, Claude Sonnet 5.5, and Claude Haiku 5.5 as the current models. All support text and image input, text output, vision, tool use, and a 1M-token context window, with a maximum output of 128K tokens.14
- The documentation suggests Claude Opus 5.5 as the default starting point and Claude Fable 5.1 for demanding reasoning and long-horizon agentic work. Claude Haiku 5.5 is aimed at high-volume tasks such as classification, extraction, and routing.14
- Anthropic added the Fable and Mythos lines in June 2026. It describes Mythos-class models as a tier above its Opus class. Claude Fable 5 was the general-release version. Claude Mythos 5 was the same underlying model with some safeguards lifted, for vetted partners.35
- Claude Fable 5.1 and Claude Mythos 5.1, released September 1, 2026, are the same model with different safeguard levels. Mythos 5.1's safeguards are tuned for cybersecurity and life sciences work, and only verified organizations can use it.3421
- Current models use "adaptive thinking": the model decides how much to reason before answering, guided by an effort setting. On Claude Fable 5.1 and Claude Opus 5.5 this thinking is always on.14
- Anthropic sorts models into four lifecycle stages: Active, Legacy, Deprecated, and Retired. It gives at least 60 days' notice before retiring a publicly released model. Its retirement dates apply to the platforms Anthropic operates; Amazon Bedrock and Google Cloud set their own.19
- Legacy models that are still available include Claude Fable 5, Claude Opus 5, Claude Opus 4.8, 4.7, 4.6, and 4.5, Claude Sonnet 5, Claude Sonnet 4.6, and Claude Haiku 4.5.14
- Recent retirements include Claude Haiku 3.5 and Claude Sonnet 3.7 (February 19, 2026), Claude Haiku 3 (April 20, 2026), Claude Opus 4 and Claude Sonnet 4 (June 15, 2026), and Claude Opus 4.1 (August 5, 2026). Claude Sonnet 4.5 was deprecated on September 30, 2026 and is due to retire on November 30, 2026.19
- Claude Mythos Preview was offered from April 2026 through Project Glasswing, an effort to find and fix flaws in critical software. Anthropic said it did not plan to make the model generally available, and it was deprecated on June 9, 2026.3819
- Earlier generations were Claude and Claude Instant (March 2023), Claude 2 (July 2023), Claude 2.1 (November 2023), the Claude 3 family (March 2024), Claude 3.5 Sonnet (June 2024), Claude 3.7 Sonnet (February 2025), and Claude Opus 4 and Claude Sonnet 4 (May 2025). All of these have since been retired.4847464543423919
- In November 2025 Anthropic committed to keep the weights of all publicly released models, and of models used significantly inside the company, for at least as long as the company exists. It also said it would interview models about their development before deprecating them.53
Claude apps, plans, and features
The Claude app runs on the web, on desktop (macOS, Windows, and a Linux beta for Ubuntu and Debian), and on iOS and Android.60773
- Plans are Free, Pro, Max (with two usage tiers), Team (Standard and Premium seats), Enterprise, and Education. Team and Enterprise add admin tools such as single sign-on, and Enterprise adds features such as audit logs, SCIM, and a Compliance API.12
- App features include Projects, Artifacts, memory, connectors (including custom MCP connectors), skills, plugins, web search, voice, scheduled tasks, and file creation.712
- Research is a mode on paid plans in which Claude runs a series of searches and returns an answer with citations. It can search the web and connected Google Workspace apps such as Gmail, Google Calendar, and Google Docs.61
- On July 10, 2026 memory changed from a daily summary to separate, labeled entries that Claude updates during conversations.62
- Claude Cowork brings Claude Code's agentic abilities to non-coding work, such as organizing files and producing spreadsheets and slides. Anthropic's January 13, 2026 post introducing Labs says Cowork launched as a research preview the previous day. On September 16, 2026 Anthropic began merging Cowork and chat into one Claude experience, starting with Pro and Max.636564
- Claude Design, a product for creating designs, prototypes, and slides, launched as an Anthropic Labs research preview on April 17, 2026. Since September 16, 2026, Claude Design, Claude Slides, and Claude Docs work inside any conversation as Artifacts.666562
- Claude in Chrome is a browser extension that lets Claude read pages, click, and fill in forms. It began as a pilot in August 2025 and became generally available on every paid plan on August 26, 2026. It does not yet run on mobile or on other Chromium browsers.6768
- Claude for Microsoft 365 puts Claude inside Office apps. The Excel, PowerPoint, and Word add-ins are generally available on all paid plans, and the Outlook add-in is in beta.69
- Claude Science, launched in beta on June 30, 2026, is described by Anthropic as "an AI workbench for scientists." It bundles common research tools and keeps the code behind each figure.70
- Claude Tag, a beta for Team and Enterprise plans launched June 23, 2026, lets anyone in a Slack channel hand tasks to Claude by tagging @Claude. It replaces the earlier Claude in Slack app.71
- Claude Security, formerly a research preview called Claude Code Security, entered public beta for Enterprise customers on April 30, 2026. It scans codebases for security flaws and suggests patches.72
Claude Code, Agent SDK, and the Claude Platform
Claude Code is an agentic coding tool. It reads a codebase, edits files, runs commands, and works with developer tools from the terminal, IDEs, a desktop app, and the browser.73
- Claude Code launched as a limited research preview on February 24, 2025 and became generally available on May 22, 2025. It now also runs in VS Code, JetBrains IDEs, Slack, GitHub Actions, and GitLab CI/CD.423973
- Claude Code surfaces generally require a Claude subscription or a Claude Console account. The terminal tool, VS Code, and JetBrains can also use third-party cloud providers.73
- The Claude Agent SDK, released with Claude Sonnet 4.5 on September 29, 2025, offers the same tools, agent loop, and context management that power Claude Code, for Python and TypeScript. Its use is governed by Anthropic's Commercial Terms of Service.4074
- Claude Managed Agents, in beta, is a hosted agent harness. Sessions run in an Anthropic-managed cloud sandbox or in a self-hosted sandbox on the customer's own infrastructure.75
- The Claude Platform API groups its features into model capabilities, tools, tool infrastructure, context management, and files. Examples include batch processing, citations, prompt caching, token counting, structured outputs, and the Files API.76
- Built-in tools include web search, web fetch, and code execution, which run on Anthropic's servers, and bash, text editor, memory, computer use, and browser use, which run in the developer's environment. Computer use left beta and a browser use tool launched on August 19, 2026.7679
- Agent Skills package instructions and scripts, with pre-built skills for PowerPoint, Excel, Word, and PDF. A beta MCP connector lets the Messages API call remote MCP servers directly.76
- Besides Anthropic's own API, Claude models are on Amazon Bedrock (run by AWS), Google Cloud (run by Google), Microsoft Foundry (run by Anthropic on Azure), and Claude Platform on AWS (run by Anthropic, with AWS sign-in and billing). Feature availability differs by platform.767778
- On Microsoft Foundry, customers choose between "Hosted on Azure," where inference runs on Azure infrastructure, and "Hosted on Anthropic," which offers all Claude models available on Foundry.78
Research, safety, and policies
Anthropic's Responsible Scaling Policy (RSP) sets capability thresholds and the safeguards required when models reach them. Version 3.4 took effect on July 8, 2026.49
- RSP versions and effective dates: 1.0 (September 19, 2023), 2.0 (October 15, 2024), 2.1 (March 31, 2025), 2.2 (May 14, 2025), 3.0 (February 24, 2026), 3.1 (April 2, 2026), 3.2 (April 29, 2026), 3.3 (May 26, 2026), and 3.4 (July 8, 2026). Version 3.0 was a full rewrite that added Frontier Safety Roadmaps and Risk Reports.49
- On May 22, 2025 Anthropic turned on AI Safety Level 3 (ASL-3) protections for Claude Opus 4. These included classifiers that screen inputs and outputs for chemical, biological, radiological, and nuclear weapons content, plus security controls meant to prevent theft of model weights.51
- The Transparency Hub's Model Report summarizes each model's capabilities, safety evaluations, and safeguards, and links to its system card. It states that Claude Opus 5, Claude Opus 5.5, Claude Fable 5.1, and Claude Mythos 5.1 were deployed with ASL-3 protections.50
- Under RSP 3.0, Anthropic publishes Risk Reports covering all of its deployed models. Redacted reports are dated February 2026 and August 2026; the August report covers the period up to July 15, 2026.49
- On Claude Fable models, classifiers detect certain cybersecurity, biology, chemistry, and model-distillation requests and hand them to another Claude model. Anthropic requires 30-day data retention for traffic on Mythos-class models and says this data is used for safety, not training.35
- Since September 1, 2026, text generated by Claude Fable 5.1 and Claude Mythos 5.1 carries an Anthropic watermark, and media files from code execution carry C2PA Content Credentials. Anthropic ties the watermark to EU AI Act compliance and is previewing a detection API privately.7934
- On October 8, 2026 Anthropic announced Usage Policy changes taking effect November 12, 2026. They include a new section on deceptive campaigns, wider weapons rules that cover arming drones, and requirements for models that control hardware able to act physically.8716
- On January 22, 2026 Anthropic published a new constitution for Claude, describing the values and behavior it intends for Claude. It is released under a CC0 1.0 public domain dedication.52
- The Anthropic Economic Index, launched February 10, 2025, studies how AI use affects work and the economy, and its initial dataset was published on Hugging Face. 2026 reports include "Learning curves" (March 24) and "Cadences" (June 26).5455
- Anthropic released Petri, an open-source tool for testing models for behaviors such as deception and sycophancy. On May 7, 2026 it handed development to Meridian Labs, an AI evaluation nonprofit.57
- Anthropic's privacy policy names Anthropic PBC as the company behind its services and Anthropic Ireland, Limited as data controller for people in the EEA, the UK, and Switzerland.3
Infrastructure and compute
Anthropic says it trains and runs Claude on AWS Trainium, Google TPUs, and NVIDIA GPUs, and picks the chip that suits each workload.93
- On April 20, 2026 Anthropic and Amazon agreed on up to 5 gigawatts of capacity for training and running Claude, built on Amazon's Trainium and Graviton chips. Anthropic calls AWS its primary training and cloud provider for mission-critical workloads. The two companies built the Project Rainier compute cluster together.94
- The same Amazon agreement covered more inference capacity in Asia and Europe and bringing the full Claude Platform into AWS.94
- On April 6, 2026 Anthropic signed an agreement with Google and Broadcom for multiple gigawatts of next-generation TPU capacity, expected to come online from 2027. The bulk of it will be in the United States.93
- On May 6, 2026 Anthropic agreed to use all of the compute capacity at SpaceX's Colossus 1 data center, over 300 megawatts. The same post doubled Claude Code's five-hour usage limits on Pro, Max, Team, and seat-based Enterprise plans.95
- Claude Platform on AWS runs on infrastructure that Anthropic manages. For workspaces created from September 18, 2026 on, requests, model inference, and stored workspace content are processed on AWS infrastructure.77
- Developers can use the inference_geo setting to choose where inference runs, either "global" or "us."76
- When it announced its Sydney office in March 2026, Anthropic said it was exploring local compute capacity in Australia through third-party partners.100
Corporate structure and governance
Anthropic PBC is a Delaware public benefit corporation. The Long-Term Benefit Trust works alongside that status to tie the company's governance to its mission.21
- The board of directors is elected by stockholders and by the Long-Term Benefit Trust. The company page lists Dario Amodei, Daniela Amodei, Yasmin Razavi, Reed Hastings, Chris Liddell, and Vas Narasimhan as directors.1
- When it was announced in 2023, the Trust was set up as an independent body of trustees with no financial interest in Anthropic. It holds Class T stock, which lets it elect and remove some directors, a share planned to grow to a board majority.2
- The company page lists the current trustees of the Long-Term Benefit Trust as Neil Buddy Shah, Richard Fontaine, and Ben Bernanke.1
- Anthropic Ireland, Limited acts as data controller for the EEA, the UK, and Switzerland.3
- Country leaders named in Anthropic announcements include Hidetoshi Tojo (Representative Director and President of Anthropic in Japan), KiYoung Choi (Representative Director of Korea), Irina Ghose (Managing Director of India), and Chris Ciauri (Managing Director of International).97989996
- Anthropic has handed some projects to outside stewards: the Model Context Protocol to the Agentic AI Foundation in December 2025, and the Petri auditing tool to Meridian Labs in May 2026.5957
Government and national security
Anthropic makes custom Claude Gov models for U.S. national security customers. Access is limited to those who work in classified environments.80
- On August 12, 2025 Anthropic offered Claude for Enterprise and Claude for Government to all three branches of the U.S. government for one year through the GSA OneGov program. It said models offered through Claude for Government are certified for FedRAMP High.81
- On February 27, 2026 the General Services Administration (GSA) removed Anthropic from USAi.gov and its Multiple Award Schedule, citing a presidential directive to stop using Anthropic's technology.82
- On March 3, 2026 the Secretary of War made a determination under the Federal Acquisition Supply Chain Security Act to exclude Claude from the Department of War's supply chain. According to the D.C. Circuit, this followed Anthropic's refusal to relax contract terms that barred using Claude for lethal autonomous warfare or domestic surveillance.85
- On March 26, 2026 the U.S. District Court for the Northern District of California (Case No. 26-cv-01996-RFL) issued a preliminary injunction blocking the government from carrying out the directive. On April 3 GSA said it would restore Anthropic's earlier status.83
- According to the D.C. Circuit's opinion, the Northern District of California decided its case on August 27, 2026 and set aside the department's designation of Anthropic as a "supply chain risk" under 10 U.S.C. § 3252.85
- On September 25, 2026 the U.S. Court of Appeals for the D.C. Circuit (No. 26-1049) denied Anthropic's petitions for review of the separate exclusion under 41 U.S.C. § 4713. One judge dissented.85
- As of October 11, 2026, GSA's artificial intelligence page carried an alert that GSA will remove any system integrations with Anthropic by August 27, 2026. The page did not give a reason.84
- On October 8, 2026 Anthropic announced commitments to the federal Genesis Mission. Over three years it plans to provide Claude, Claude Code, and API credits to research projects and to work with agencies and national laboratories.86
- Anthropic says it developed the Life Sciences Verification Program with the U.S. government. Organizations seeking the Cyber Verification Program's Specialized Access tier are reviewed in depth with the U.S. government.3489
- Outside the U.S., Anthropic signed a Memorandum of Cooperation on AI evaluation methods with the Japan AI Safety Institute (October 2025). It also signed a Memorandum of Understanding on AI safety and cybersecurity with Korea's Ministry of Science and ICT (June 2026).9798
Offices and international presence
Anthropic is headquartered in San Francisco, California, and has announced offices across North America, Europe, and Asia-Pacific.396
- A November 7, 2025 announcement named offices in San Francisco, Seattle, New York, Washington D.C., London, Dublin, Zurich, Tokyo, Seoul, and Bengaluru, and announced new offices in Paris and Munich.96
- Anthropic opened a Tokyo office in October 2025.97
- Anthropic opened a Bengaluru office on February 16, 2026, led by Irina Ghose. The announcement described partnerships, including a government MCP server from India's Ministry of Statistics.99
- Anthropic opened a Seoul office in June 2026, led by KiYoung Choi.98
- On March 10, 2026 Anthropic announced a Sydney office, its fourth in Asia-Pacific, focused at the start on enterprise, startup, and research customers.100
Recent updates (July to October 2026)
- July 1, 2026: Access to Claude Fable 5 and Claude Mythos 5 was restored after being suspended on June 12. On July 8, version 3.4 of the Responsible Scaling Policy took effect.357949
- July 24, 2026: Anthropic released Claude Opus 5. With it, the API gained beta support for changing tools in the middle of a conversation and for automatic fallback to another model when a request is flagged.36
- August 2026: Anthropic published a redacted Risk Report covering the period up to July 15. On August 19, computer use left beta, a browser use tool launched, and the Files API and Agent Skills left beta.4979
- August 26, 2026: Claude in Chrome became generally available on every paid plan.68
- August 27, 2026: Anthropic opened a research preview of the Model Hardware Standard, a shared specification for AI agents to safely operate lab instruments and other physical devices. Anthropic says it plans to make the standard open source later.92
- September 1, 2026: Anthropic released Claude Fable 5.1 and Claude Mythos 5.1. It also announced Enterprise Frontier Safeguards, which pair zero data retention with misuse detection on cloud infrastructure the customer controls. Rollout is in phases starting in fall 2026.3491
- September 16, 2026: Anthropic began merging Claude Cowork and chat into one Claude, and made Claude Design, Claude Slides, and Claude Docs available inside conversations.65
- September 17, 2026: Anthropic launched the Life Sciences Verification Program in beta. On October 6 it expanded the Cyber Verification Program into three tiers (Defense, Red Team, and Specialized Access).9089
- September 22 to October 7, 2026: Anthropic released Claude Opus 5.5 (September 22), Claude Sonnet 5.5 (September 28), and Claude Haiku 5.5 (October 7). On September 30 it announced the deprecation of Claude Sonnet 4.5.31323319
- September 25, 2026: The D.C. Circuit denied Anthropic's petitions for review of the Department of War's supply chain exclusion.85
- October 7, 2026: Python and TypeScript SDKs added beta tool classes for computer use and browser use, and Max and Team plans began to include monthly API credits.79
- October 8, 2026: Anthropic announced Usage Policy updates (effective November 12), the Anthropic Cyber Mission, and commitments to the Genesis Mission. The Cyber Mission includes a defense program for critical infrastructure and an opt-in security scanner for open-source projects.878886
Models and products lineup
Newest first. Status is as the official source describes it; “Legacy” and “Retired” entries stay listed for history.
| Name | Kind | Status | Released | What it is |
|---|---|---|---|---|
| Claude Haiku 5.5 (external site: anthropic.com) | Model | Current | Oct 7, 2026 | Small model for high-volume, cost-sensitive tasks such as classification, extraction, and routing, with a 1M-token context window. Offered on the Claude Platform (claude-haiku-5-5) and through AWS, Google Cloud, and Microsoft Azure.3314 |
| Claude Sonnet 5.5 (external site: anthropic.com) | Model | Current | Sep 28, 2026 | Faster, lower-cost companion to Claude Opus 5.5 for well-scoped everyday tasks, with a 1M-token context window. Offered on the Claude Platform (claude-sonnet-5-5) and through AWS, Google Cloud, and Microsoft Azure.3214 |
| Claude Opus 5.5 (external site: anthropic.com) | Model | Current | Sep 22, 2026 | Opus model for long-running agentic coding and knowledge work, with a 1M-token context window. Offered in Claude plans, Claude Code, and the Claude Platform (claude-opus-5-5), and through AWS, Google Cloud, and Microsoft Azure.3114 |
| Claude Fable 5.1 (external site: platform.claude.com) | Model | Current | Sep 1, 2026 | Model for demanding reasoning and long-horizon agentic work, with a 1M-token context window and always-on adaptive thinking. Offered on the Claude API (claude-fable-5-1), Amazon Bedrock, Google Cloud, Microsoft Foundry, and Claude Platform on AWS.2034 |
| Claude Mythos 5.1 (external site: platform.claude.com) | Model | Limited access | Sep 1, 2026 | The same model as Claude Fable 5.1, with safeguards tuned for cybersecurity and life sciences work. Only organizations verified through Anthropic's verification programs, such as the Cyber Verification Program, can use it.2134 |
| Claude Opus 5 (external site: platform.claude.com) | Model | Legacy | Jul 24, 2026 | Opus model with a 1M-token context window, now marked legacy but still available. Anthropic recommends moving to Claude Opus 5.5.2336 |
| Claude Sonnet 5 (external site: platform.claude.com) | Model | Legacy | Jun 30, 2026 | Sonnet model that Anthropic described as built for agentic work such as planning and using browsers and terminals. Now legacy but still available; Claude Sonnet 5.5 is the current Sonnet.2437 |
| Claude Science (external site: anthropic.com) | App | Preview | Jun 30, 2026 | Beta app for macOS and Linux that Anthropic calls "an AI workbench for scientists." It is offered on Pro, Max, Team, and Enterprise plans.70 |
| Claude Fable 5 (external site: platform.claude.com) | Model | Legacy | Jun 9, 2026 | Anthropic called it a Mythos-class model made safe for general use. Access was suspended from June 12 to July 1, 2026. Now legacy but still available; Claude Fable 5.1 is the current Fable model.2235 |
| Claude Mythos 5 | Model | Limited access | Jun 9, 2026 | The same underlying model as Claude Fable 5 with some safeguards lifted. It was offered at launch to Project Glasswing partners and a small group of cyber defenders and infrastructure providers, with access for selected biology researchers planned. Platform listings mark it as limited availability.351978 |
| Claude Opus 4.8 (external site: platform.claude.com) | Model | Legacy | May 28, 2026 | Opus model with a 1M-token context window, now legacy but still available on the Claude API and cloud platforms. Retirement is not sooner than May 28, 2027.25 |
| Claude Security (external site: claude.com) | Service | Preview | Apr 30, 2026 | Scans codebases for vulnerabilities and generates targeted patches. Public beta for Enterprise customers; formerly the Claude Code Security research preview.72 |
| Claude Opus 4.7 (external site: platform.claude.com) | Model | Legacy | Apr 16, 2026 | Opus model that introduced the tokenizer used by current models. Now legacy but still available; retirement is not sooner than April 16, 2027.26 |
| Claude Mythos Preview (external site: anthropic.com) | Model | Legacy | Apr 7, 2026 | General-purpose model offered to Project Glasswing partners and other organizations that maintain critical software, to find and fix vulnerabilities. Anthropic did not make it generally available, and it was deprecated on June 9, 2026.3819 |
| Claude Sonnet 4.6 (external site: platform.claude.com) | Model | Legacy | Feb 17, 2026 | Sonnet model with a 1M-token context window, now legacy but still available. Retirement is not sooner than February 17, 2027.28 |
| Claude Opus 4.6 (external site: platform.claude.com) | Model | Legacy | Feb 5, 2026 | Opus model with a 1M-token context window, now legacy but still available. Retirement is not sooner than February 5, 2027.27 |
| Claude Opus 4.5 (external site: platform.claude.com) | Model | Legacy | Nov 24, 2025 | Opus model with a 200K-token context window, now legacy but still available. Retirement is not sooner than November 24, 2026.2919 |
| Claude Haiku 4.5 (external site: platform.claude.com) | Model | Legacy | Oct 15, 2025 | Small model with a 200K-token context window, now legacy but still available. Retirement is not sooner than October 15, 2026.3019 |
| Claude Sonnet 4.5 (external site: anthropic.com) | Model | Legacy | Sep 29, 2025 | Sonnet model released under ASL-3 protections. Deprecated on September 30, 2026, with retirement on the Claude API set for November 30, 2026.4019 |
| Claude Agent SDK (external site: code.claude.com) | API | Current | Sep 29, 2025 | Python and TypeScript libraries that give developers the tools, agent loop, and context management behind Claude Code, for building their own agents.7440 |
| Claude in Chrome (external site: claude.com) | App | Current | Aug 25, 2025 | Chrome extension that lets Claude read, click, and fill in forms on web pages. It began as a pilot in August 2025 and became generally available on every paid plan on August 26, 2026.6768 |
| Claude Opus 4.1 (external site: anthropic.com) | Model | Retired | Aug 5, 2025 | Upgrade to Claude Opus 4 for agentic tasks, coding, and reasoning. Retired on August 5, 2026.4119 |
| Claude Opus 4 (external site: anthropic.com) | Model | Retired | May 22, 2025 | Hybrid model with near-instant replies and an extended thinking mode. It was the model for which Anthropic turned on ASL-3 protections. Retired on June 15, 2026.395119 |
| Claude Sonnet 4 (external site: anthropic.com) | Model | Retired | May 22, 2025 | Sonnet model released alongside Claude Opus 4 as an upgrade to Claude Sonnet 3.7. Retired on June 15, 2026.3919 |
| Claude Gov (external site: anthropic.com) | Model | Limited access | Mar 2025 | Custom Claude models built for U.S. national security customers. Access is limited to those who work in classified environments.8085 |
| Claude 3.7 Sonnet (external site: anthropic.com) | Model | Retired | Feb 24, 2025 | Hybrid reasoning model that could answer quickly or use visible step-by-step "extended thinking." Retired on February 19, 2026.4219 |
| Claude Code (external site: code.claude.com) | Agent | Current | Feb 24, 2025 | Agentic coding tool for the terminal, IDEs, desktop, web, and mobile. It is included in Claude subscription plans or used through a Claude Console account.7342 |
| Model Context Protocol (MCP) (external site: modelcontextprotocol.io) | Platform | Current | Nov 25, 2024 | Open standard Anthropic introduced for connecting AI applications to external data sources and tools. Anthropic donated it to the Agentic AI Foundation in December 2025.5859 |
| Claude 3.5 Haiku (external site: anthropic.com) | Model | Retired | Oct 22, 2024 | Small model announced together with the upgraded Claude 3.5 Sonnet and the computer use beta. Retired on February 19, 2026.4419 |
| Claude 3.5 Sonnet (external site: anthropic.com) | Model | Retired | Jun 2024 | Mid-sized model launched with the Artifacts preview in Claude.ai, with an upgraded version on October 22, 2024. Both versions were retired on October 28, 2025.434419 |
| Claude 3 Opus (external site: anthropic.com) | Model | Retired | Mar 4, 2024 | Large model in the Claude 3 family, announced with Claude 3 Sonnet and Claude 3 Haiku, with a 200K-token context window at launch. Retired on January 5, 2026; Claude 3 Haiku followed on April 20, 2026.4519 |
| Claude 2 (external site: anthropic.com) | Model | Retired | Jul 11, 2023 | Model offered through the API and the then-new claude.ai beta website. Claude 2 and Claude 2.1 were retired on July 21, 2025.4719 |
| Claude app (external site: claude.com) | App | Current | Jul 11, 2023 | Claude's own app on the web (claude.ai), desktop, and iOS and Android, with Free, Pro, Max, Team, Enterprise, and Education plans. Claude.ai opened as a public beta website with Claude 2.6012747 |
| Claude Instant (external site: anthropic.com) | Model | Retired | Mar 14, 2023 | A lighter, lower-cost, quicker option introduced alongside the original Claude. Instant versions 1.0 to 1.2 were retired on November 6, 2024.4819 |
| Claude (external site: anthropic.com) | Model | Retired | Mar 14, 2023 | Anthropic's original AI assistant model, offered through a chat interface and an API. Claude 1.0 to 1.3 were retired on November 6, 2024.4819 |
| Claude Platform (Claude API) (external site: platform.claude.com) | API | Current | Mar 2023 | Anthropic's API for Claude models, with keys issued in the Claude Console. Anthropic has offered Claude through an API since March 2023.87648 |
| Claude Managed Agents (external site: platform.claude.com) | Service | Preview | Not stated | Beta hosted agent harness on the Claude Platform. Claude sessions run in an Anthropic-managed cloud sandbox or a self-hosted sandbox, with built-in tools and MCP servers.75 |
| Claude Cowork (external site: claude.com) | Agent | Current | Not stated | Brings Claude Code's agentic abilities to knowledge work beyond coding, on paid plans. Since September 16, 2026 it has been merging into the main Claude app.6365 |
| Claude for Microsoft 365 (external site: claude.com) | App | Current | Not stated | Add-ins that put Claude inside Excel, PowerPoint, and Word (generally available on paid plans) and Outlook (beta).69 |
Timeline
Mar 14, 2023
Anthropic introduced Claude and Claude Instant through a chat interface and an API.48
Jul 11, 2023
Anthropic released Claude 2 and opened the claude.ai beta website in the US and UK.47
Sep 19, 2023
Anthropic announced the Long-Term Benefit Trust, and version 1.0 of its Responsible Scaling Policy took effect.249
Nov 21, 2023
Claude 2.1 added a 200K-token context window, system prompts, and beta tool use.46
Mar 4, 2024
Anthropic introduced the Claude 3 family (Claude 3 Opus, Claude 3 Sonnet, and Claude 3 Haiku).45
Jun 2024
Anthropic released Claude 3.5 Sonnet and a preview of Artifacts on Claude.ai.43
Oct 15, 2024
Version 2.0 of the Responsible Scaling Policy took effect.49
Oct 22, 2024
Anthropic released an upgraded Claude 3.5 Sonnet and Claude 3.5 Haiku, and opened a public beta of computer use.44
Nov 25, 2024
Anthropic introduced the Model Context Protocol as an open standard.58
Feb 10, 2025
Anthropic launched the Anthropic Economic Index and published its dataset on Hugging Face.54
Feb 24, 2025
Anthropic released Claude 3.7 Sonnet and a limited research preview of Claude Code.42
May 22, 2025
Anthropic released Claude Opus 4 and Claude Sonnet 4, made Claude Code generally available, and turned on ASL-3 protections for Claude Opus 4.3951
Jun 6, 2025
Anthropic announced Claude Gov models for U.S. national security customers.80
Aug 12, 2025
Anthropic offered Claude for Enterprise and Claude for Government to all three branches of the U.S. government through GSA's OneGov program.81
Aug 25, 2025
Anthropic began a pilot of Claude for Chrome, a browser extension.67
Sep 29, 2025
Anthropic released Claude Sonnet 4.5 and the Claude Agent SDK.40
Nov 4, 2025
Anthropic published commitments to preserve the weights of publicly released models.53
Nov 24, 2025
Anthropic released Claude Opus 4.5.29
Dec 9, 2025
Anthropic donated the Model Context Protocol to the Agentic AI Foundation under the Linux Foundation.59
Jan 22, 2026
Anthropic published a new constitution for Claude under a CC0 public domain dedication.52
Feb 2026
Anthropic released Claude Opus 4.6 (February 5) and Claude Sonnet 4.6 (February 17).2728
Feb 24, 2026
Version 3.0 of the Responsible Scaling Policy, a full rewrite, took effect.49
Feb 27, 2026
GSA removed Anthropic from USAi.gov and its Multiple Award Schedule, citing a presidential directive.82
Mar 3, 2026
The Secretary of War made a determination to exclude Claude from the Department of War's supply chain.85
Mar 26, 2026
A federal district court in Northern California issued a preliminary injunction against the directive, and GSA later restored Anthropic's earlier status.83
Apr 7, 2026
Anthropic announced Project Glasswing and gave partners access to Claude Mythos Preview.38
Apr 16, 2026
Anthropic released Claude Opus 4.7.26
May 28, 2026
Anthropic released Claude Opus 4.8.25
Jun 9, 2026
Anthropic released Claude Fable 5 and Claude Mythos 5.35
Jun 30, 2026
Anthropic released Claude Sonnet 5 and a beta of Claude Science.3770
Jul 24, 2026
Anthropic released Claude Opus 5.36
Aug 26, 2026
Claude in Chrome became generally available on every paid plan.68
Aug 27, 2026
The Northern District of California set aside the Department of War's designation of Anthropic as a supply chain risk under 10 U.S.C. § 3252.85
Sep 1, 2026
Anthropic released Claude Fable 5.1 and Claude Mythos 5.1.3420
Sep 16, 2026
Anthropic began merging Claude Cowork and chat into one Claude experience.65
Sep 22, 2026
Anthropic released Claude Opus 5.5.31
Sep 25, 2026
The D.C. Circuit denied Anthropic's petitions for review of its exclusion under 41 U.S.C. § 4713.85
Sep 28, 2026
Anthropic released Claude Sonnet 5.5.32
Oct 7, 2026
Anthropic released Claude Haiku 5.5.33
Oct 8, 2026
Anthropic announced Usage Policy updates, the Anthropic Cyber Mission, and commitments to the Genesis Mission.878886
Leadership
Titles as the official source lists them.
Dario Amodei
Co-Founder and Chief Executive Officer18
Daniela Amodei
Co-Founder and President18
Tom Brown
Co-Founder and Chief Compute Officer18
Jack Clark
Co-Founder and Head of Public Benefit18
Jared Kaplan
Co-Founder and Chief Science Officer18
Sam McCandlish
Co-Founder and Chief Architect18
Chris Olah
Co-Founder and Interpretability Research Lead18
Mariano-Florentino (Tino) Cuéllar
Chief Global Affairs Officer18
Rahul Patil
Chief Technology Officer18
Krishna Rao
Chief Financial Officer18
Paul Smith
Chief Commercial Officer18
Ami Vora
Chief Product Officer18
Using this organization’s offerings
Intended users
Anthropic sells Claude plans for individuals (Free, Pro, and Max) and for teams and enterprises, and documents the Claude API for developers building Claude into applications and agents.1213
Hosted and self-hosted access
Claude models are offered as hosted services: in the Claude apps, through the Claude API with keys issued in the Claude Console, and on cloud platforms including Amazon Bedrock, Google Cloud, and Microsoft Foundry.131412
Restrictions and review limits
Anthropic's Hugging Face organization lists datasets but no public models. Use is governed by Anthropic's Usage Policy; Claude Mythos models are available only to organizations verified through Anthropic's verification programs, and Amazon Bedrock and Google Cloud set their own pricing and retirement dates for Claude models. This profile is not a complete product or rights audit.17161514218990
Key facts
- Ownership (separate from openness)
- Privately held
Recorded classification; fact-level ownership verification pending.
- Other documented locations
- Dublin, Ireland (Anthropic Ireland, Limited)3
- Founded
- Unknown
- Sectors
- Frontier models, Agents, Research
What is and isn’t public
Products
Hosted model products
Labeled Hosted product/API — this describes how a documented product is delivered (hosted access). It does not determine whether underlying weights or code are open; assess the specific release and its licenses.
| Product | Kind | Documented delivery and access | Reviewed |
|---|---|---|---|
| Claude (external site: claude.com) | Assistant app | AI assistant on the web and in desktop (macOS, Windows, and a Linux beta) and mobile apps, with Free, Pro, Max, Team, Enterprise, and Education plans; Claude Cowork is being merged into the main app from September 2026.766065 | Oct 11, 2026 |
| Claude Platform (Claude API) (external site: platform.claude.com) | Hosted model API | API access to Claude models with API keys through the Claude Console; Claude is also offered on Amazon Bedrock, Google Cloud, Microsoft Foundry, and Claude Platform on AWS (operated by Anthropic, with AWS sign-in and billing).87677 | Oct 11, 2026 |
Other products
| Product | Kind | Documented delivery and access | Reviewed |
|---|---|---|---|
| Claude Code (external site: claude.com) | Developer tool | Agentic coding tool for the terminal, IDEs, desktop, web, and mobile; included in Claude subscription plans or billed through a Claude Console account.9 | Sep 29, 2026 |
Cataloged artifacts
Frameworks
- Model Context Protocol (MCP)SoftwareAvailability: Public
Evaluation tools
- Petri (Inspect Petri)Evaluation toolAvailability: Public
Notable documented facts
In the news
Structured outputs for Anthropic Claude models become generally available on Amazon Bedrock
In an October 19, 2026 post on the AWS Machine Learning Blog, AWS said structured outputs for Claude on Amazon Bedrock is now generally available for all Anthropic Claude models, on both the bedrock-runtime and bedrock-mantle endpoints. AWS says the feature uses constrained decoding to enforce a developer's JSON schema during token generation, and that setting
strict: trueon a tool constrains the model's tool input to match its schema. The post also says some JSON Schema keywords, including numeric and string length constraints, are accepted but not enforced, so applications should validate those values themselves.Anthropic gives a METR team employee-like access and describes publication terms for outside evaluators
In an October 15, 2026 update on independent evaluators, Anthropic said it has begun giving a small team from METR employee-like access, with internal tools and systems roughly comparable to those of its own risk assessment teams, under enforceable agreements, and that METR will publish its risk assessments. Anthropic says it cannot veto publication but can require redactions limited to narrow, predefined categories such as customer data, privileged information, and model architectures, and that evaluators may say publicly whether redactions were justified. The post describes three tiers of evaluator access and says SecureBio is reviewing the chemical and biological weapons chapter of Anthropic's August Risk Report. These are Anthropic's own descriptions; the post does not include METR's assessments.
Anthropic describes Claude election safeguards ahead of the 2026 U.S. midterms
On October 14, 2026 Anthropic described its preparations for the 2026 U.S. midterm elections. It says Claude.ai continues to show a banner linking to TurboVote from Democracy Works for questions about registration, polling places, and election dates, that later this month it will draw official voting information from Democracy Works into responses in the United States, and that Claude will show up-to-date results on election night and in the following days. Anthropic also says it updated its Usage Policy's election rules, published a guide for campaigns, and added classifiers aimed at multi-step abuse such as fake social media accounts and news sites. It says the Institute for Strategic Dialogue and the McCain Institute tested these safeguards and found Claude could miss a malicious campaign spread across many harmless-looking requests.
Anthropic opens a Singapore office and signs an MOU with Singapore's digital ministry
On October 14, 2026 Anthropic announced that it has opened an office in Singapore, which it describes as its fifth in the Asia-Pacific region. Anthropic co-founder Ben Mann and Kai Fong Chng, Permanent Secretary of Singapore's Ministry of Digital Development and Information, signed a memorandum of understanding on the sidelines of Singapore International Cyber Week. The announcement says the MOU covers cooperation on AI safety, security, and model evaluation, research on AI adoption and its effects, and responsible adoption across the public and private sectors, including technical exchanges with Singapore's AI safety and assurance institutions. An MOU sets out intended cooperation; the announcement does not describe binding commitments.
Anthropic adds dynamic workflows to Claude Managed Agents in beta
Anthropic's Claude Platform release notes, in an entry dated October 9, 2026, say a Claude Managed Agents agent can now use dynamic workflows, in beta with the managed-agents-2026-04-01 beta header. The notes describe a workflow as a program that runs many agents in phases and combines their results, which the server runs in the background as a workflow run. Developers turn the feature on through the agent's multiagent field and follow each run through workflow_run events on the session's event stream. This is a feature of Anthropic's hosted service, not an open-source or open-weight release, and USASI has not tested it.
What this catalog does not know
- Founding date: not documented in the sources reviewed.
Have a primary source? How to report a correction.
U.S. eligibility
Eligible · basis: U.S. headquarters
Anthropic's transparency hub describes it as a public benefit corporation headquartered in San Francisco, California. Its privacy policy describes the company as based and headquartered in the United States and lists Anthropic PBC at a San Francisco, California address. A separate Irish entity serves as data controller for the EEA, UK, and Switzerland.53
Sources
This listing is not an endorsement or a federal approval.
Share or cite this page
- X (opens X)
- Bluesky (opens Bluesky)
- LinkedIn (opens LinkedIn)
- Reddit (opens Reddit)
- Hacker News (opens Hacker News)
- Email (opens your email app)
- RSS updates
Cite this page
APA style
USASI. (2026, October 12). Anthropic. United States of America Superintelligence. https://unitedstatesofamericasuperintelligence.com/companies/anthropic/
BibTeX
@misc{usasi_anthropic,
author = {{USASI}},
title = {Anthropic},
year = {2026},
month = oct,
howpublished = {\url{https://unitedstatesofamericasuperintelligence.com/companies/anthropic/}},
note = {United States of America Superintelligence. Last updated 2026-10-12}
}The date is when this page was last updated. Add the date you read it if your style needs one.
For maintainers: a “Listed in USASI” badge
Links to this page. It says the entry is listed, not endorsed or rated. About the badge
Markdown (for a README)
[](https://unitedstatesofamericasuperintelligence.com/companies/anthropic/)
HTML
<a href="https://unitedstatesofamericasuperintelligence.com/companies/anthropic/"><img src="https://unitedstatesofamericasuperintelligence.com/badges/listed.svg" alt="Listed in the USASI catalog" height="20"></a>