Anthropic Says Claude Now Leads 26% of Its AI R&D Work

Anthropic says its AI model Claude now leads 26% of the company’s measured artificial intelligence research and development work, a sharp increase from less than 1% in February 2026. More than 90% of Anthropic’s AI R&D work now involves AI in at least a collaborative role, although the company says Claude has not reached full autonomy in any measured area.

Anthropic uses an R&D Automation Index ranging from AL0, where AI is not involved, to AL5, where AI performs work fully autonomously. Claude reached AL4, described as “leads”, for 26% of measured R&D tasks by August. At this level, AI can complete most of a task from a high-level prompt while remaining under human supervision.

The company is also tracking around 30,000 AI agents working simultaneously on its internal platform. Anthropic says it analysed more than 1 billion decisions made by research and engineering agents in August, while about 6% of AI R&D computing resources during one week in July were allocated to safety work. The company says publishing such measurements could help the industry better track how quickly AI development is changing.