AI TRENDS | Anthropic Publishes Metrics Tracking Claude’s Role in AI Research

According to Wallstreetcn, Anthropic published a set of metrics designed to help the public track AI development progress inside frontier AI labs, showing that as of August 2026 Claude led 26% of the company’s AI research work, more than 90% of tasks reached automation at or above the "AI collaboration" level, and Claude had not yet achieved fully autonomous operation in any measured subset of AI research work. In August, about 30,000 agents were working simultaneously on research and engineering inside Anthropic, while the online monitoring system intercepted about 0.002% of agent decisions, or 1 in about 47,000 decisions, during the month. In the week of July 13-20, about 6% of Anthropic’s AI research computing power was allocated to safety work, and about 12% of the computing power used for AI-driven AI research in the same week went to safety-related tasks. Anthropic said its goal is to narrow the information gap between what frontier labs know and what the public knows, and it plans to introduce independent third-party evaluators that can access internal processes and data to verify safety practices, report incidents, and monitor key metrics.