When Was Anthropic Founded? History and Timeline

Anthropic was founded in 2021 by siblings Dario Amodei and Daniela Amodei, together with colleagues who had worked at OpenAI. Dario became CEO and Daniela became president. The company was set up as a public benefit corporation and built around research into reliable, interpretable, and steerable AI systems.

That is the short answer. The fuller history runs from an OpenAI breakaway, through early safety research, to the launch of Claude and its development into one of the major AI model families.

Where did Anthropic come from?

Anthropic emerged from a group of former OpenAI researchers and leaders. Dario Amodei had led research at OpenAI, while Daniela Amodei had held leadership roles there. They left with several colleagues and formed a new company with a more concentrated research agenda around AI reliability and safety.

Anthropic's first public funding announcement came in May 2021, when the company reported a $124 million Series A. That announcement named Dario as CEO and Daniela as president and said the team would work on large-scale AI systems that were steerable, interpretable, and robust.

Why was Anthropic founded?

Anthropic's stated purpose is to build AI systems people can rely on and to study their opportunities and risks. Its company page describes the lab as an AI safety and research company focused on reliable, interpretable, and steerable systems.

That mission produced several important strands of work:

• Constitutional AI: a training method that uses a written set of principles, model-generated critiques, and AI feedback to shape behavior. Anthropic published the underlying Constitutional AI research in 2022.
• Interpretability: research intended to make the internal behavior of neural networks more understandable.
• Responsible scaling: policies that connect stronger model capabilities with additional evaluation and safeguards.

These are research methods and governance commitments, not proof that Claude cannot be jailbroken, manipulated, or produce unsafe or incorrect outputs. Anthropic evaluates and updates its safeguards as its models change.

Anthropic history: key dates

• 2021: Anthropic is founded as a public benefit corporation by Dario and Daniela Amodei with former OpenAI colleagues. It announces a $124 million Series A in May.
• 2022: Anthropic publishes its Constitutional AI work, describing a method for training an assistant with a set of written principles and AI-generated feedback.
• March 2023: Anthropic introduces Claude after a closed alpha with partners including Notion, Quora, and DuckDuckGo, and offers access by request.
• July 2023: Claude 2 launches with API access and a public-facing web beta in the US and UK, making Claude broadly available to the public.
• March 2024: Anthropic launches the Claude 3 family, using the Haiku, Sonnet, and Opus tiers to distinguish speed, balance, and capability.
• June 2024: Claude 3.5 Sonnet arrives, alongside the first version of Artifacts.
• 2025 onward: Anthropic continues releasing new Claude generations and expanding Claude's use in coding, analysis, writing, and agentic workflows.
• September 2026: magicdoor.ai's current Claude lineup includes Claude Opus 5, Claude Sonnet 5, and Claude Fable 5.1.

How Claude changed Anthropic

Before Claude, Anthropic was known mainly as a research lab. Claude turned that research into a public product. The assistant first reached outside users through partner integrations, then expanded through Anthropic's own chat interface and API.

The 2024 Claude 3 release also established the naming structure that made the model family easier to understand: Haiku for speed and cost, Sonnet for a balance of capability and efficiency, and Opus for the most demanding work. Individual models have changed since then, but the tiered approach became part of Claude's identity.

Anthropic today

Anthropic remains an AI research and product company led by Dario and Daniela Amodei. Claude is its best-known product, while interpretability, model evaluations, and scaling policy remain central parts of the company's public work.

If you want to try current Claude models, magicdoor.ai offers Claude Opus 5, Claude Sonnet 5, and Claude Fable 5.1 in the same workspace as other major AI models. You can open a chat, compare their answers, and switch models mid-conversation.

FAQ

When was Anthropic founded?

Anthropic was founded in 2021. It was created as a public benefit corporation by siblings Dario and Daniela Amodei with a group of colleagues who had worked at OpenAI.

Who founded Anthropic?

Anthropic was co-founded by siblings Dario Amodei, its CEO, and Daniela Amodei, its president, together with other former OpenAI researchers and leaders.

Why was Anthropic started?

Anthropic was started to research and build AI systems that are reliable, interpretable, and steerable. Its early work focused on AI safety, human feedback, and understanding how large models behave.

When did Anthropic launch Claude?

Anthropic introduced Claude in March 2023 after testing it with partners in a closed alpha and offered access by request. Claude became broadly available to the public in the US and UK in July 2023 through Claude 2's public-facing web beta.