Claude Glossary

Anthropic's family of large language models, commonly offered under the Opus, Sonnet, and Haiku tier names.

Claude is the family name for the large language models produced by Anthropic. The name refers to Claude Shannon, the American mathematician and electrical engineer (1916–2001) whose 1948 paper A Mathematical Theory of Communication founded information theory.

Anthropic commonly uses three tier names to describe different points on the capability, cost, and latency curve:

• Claude Opus — the high-capability tier, positioned for difficult reasoning, writing, coding, and agentic tasks.
• Claude Sonnet — the middle tier, positioned between capability, cost, and latency.
• Claude Haiku — the small, fast tier, positioned for lower-cost and latency-sensitive work.

The family began with Claude 1 in March 2023 and has since gone through several numbered generations. The precise model attached to each tier changes, so operators should treat Opus, Sonnet, and Haiku as product positions rather than fixed capability levels.

Claude was the principal cooperative-writing model for much of this Dictionary’s early corpus. That history created the dependency described in Opus Addict, even as the operating stack later expanded to other model families.

See also

• Anthropic — the producer
• Claude Opus, Claude Sonnet, Claude Haiku — the three tiers
• Opus Addict
• Constitutional AI — Anthropic’s training approach

Return to Dictionary All Entries (A–Z) For Students Other Writing Capstone 2.0