Claude now leads 26% of Anthropic's AI research and development

Anthropic's AI model Claude now leads about 26% of the company's AI research and development work, up from effectively zero in February, according to an internal review reported by The Washington Post. Claude can complete most tasks from high-level prompts with human supervision, though it is not fully autonomous in any measured work. Over 90% of R&D involves Claude collaborating with or leading researchers. Anthropic says the rise signals progress toward "recursive self-improvement," where AI develops advanced successors with limited human input. It also reported 30,000 AI agents working on its internal platform and urged other developers to publish similar measurements.