Anthropic Glossary

AI lab founded 2021 by Dario and Daniela Amodei and former OpenAI staff; makers of the Claude model family. The Dictionary's primary frontier-model provider.

Anthropic is an AI safety company founded in 2021 by Dario Amodei, Daniela Amodei, and several other former OpenAI staff who left over disagreements about the pace and orientation of frontier AI development. The lab develops the Claude family of large language models (Opus, Sonnet, Haiku) and the broader Constitutional AI training paradigm. Headquartered in San Francisco; major investors include Google, Amazon, and Salesforce.

For this Dictionary, Anthropic is the frontier-model provider the operator runs. The Opus Addict entry names the dependency on Anthropic’s most-capable tier as a structural feature of careful agentic work in 2026, with both the upside (sustained capability gains across the 4.x series) and the downside (single-vendor reliance) acknowledged. The Lab Character entry treats Anthropic as one of the worked examples of an AI lab whose stated mission and commercial trajectory are in productive but unresolved tension.

The xAI/Anthropic data centre deal announced May 7, 2026 is one of several recent structural moves the Dictionary may revisit in future entries.

See also

• Dario Amodei
• Amanda Askell
• Claude — the model family
• Constitutional AI — Anthropic’s training approach
• Opus Addict
• Lab Character
• Sovereign Compute
• Using the Empire’s Telescope

Return to Dictionary All Entries (A–Z) For Students Other Writing