What is Claude?
Claude is the name Anthropic uses both for its family of large language models and for the assistant product built on them. The first version was released in March 2023. Later generations come in several sizes that trade capability against speed and cost, accept images as well as text, call tools, and work through long documents within a large context window.
Anthropic trains Claude with a method it calls Constitutional AI, in which the model is guided by a written set of principles alongside human feedback. Claude is available as a chat application, through Anthropic’s API, and through the major cloud providers’ AI marketplaces. Anthropic also published the Model Context Protocol, an open standard for connecting assistants to tools and data sources.
Why Claude matters
Claude is one of the main hosted model options for businesses, often chosen for long-document work such as contract review and summarization of large case files. Availability through cloud marketplaces matters to companies whose data residency commitments require traffic to stay within an existing cloud account. As with any hosted model, data leaves the company’s environment for processing, so sensitive workloads need a contract review or a private AI alternative.
What to watch for
- Model names, sizes, and context limits change with each generation, so name the exact model version in evaluations and contracts.
- Data retention and training terms differ between the consumer app, team plans, the API, and cloud marketplace access.
- Claude, like every large language model, can hallucinate, so verify facts and citations before relying on its output.