Products

Claude: An AI Model or Family of Models from Anthropic

Claude is an AI model or family of models used by Anthropic in its internal research and development work. Available sources link Claude to the company’s research on research automation, agent monitoring, and AI safety and the evaluation of its impact on user well-being.

Claude is an AI model or family of models used by Anthropic in its internal research and development work.

Claude in Anthropic’s Research

According to an experimental framework published by Anthropic, Claude led 26% of its measured research and development work, under a definition referring to a system’s ability to complete most of a task from start to finish based on general direction with human oversight. Claude did not operate with full independence in any of the research and development work measured through August 2026.

Anthropic also reported that more than 90% of the measured work involved AI at a level where the system handled significant portions of the task under close human direction. These figures remain tied to Anthropic’s internal methodology and do not represent a standardized global benchmark.

Monitoring and Safety

Anthropic stated that approximately 30,000 agents were carrying out research and engineering work on its most widely used internal platform during August 2026, with processes passing through pre-execution monitoring and subsequent monitoring to analyze behavior. This system is associated with the use of the company’s models, including Claude, in research and engineering work.

Anthropic also allocated a $5 million grants program to fund independent, open-source research measuring the impact of AI systems on user well-being. The program focuses on multi-turn conversations, mental health crises, and the risks of over-response or over-refusal.

Limits of the Available Information

  • The available sources do not include comprehensive technical specifications for Claude.
  • The sources do not provide a complete list of models, versions, or products associated with the name.
  • The available evidence is insufficient to determine pricing, subscription tiers, or a stable comparative ranking for Claude.
  • Figures concerning Claude’s use within Anthropic result from internal and experimental measurements and may change with different methodologies and periods.