Anthropic disclosed that its AI assistant Claude now leads approximately 26% of the company's internal AI research and development, up from near zero at the start of the year. The model is currently involved in about 90% of employee day-to-day R&D collaboration, with more than 30,000 agents actively executing research and engineering tasks on the platform at any given moment. To monitor this rapid integration, Anthropic plans to embed independent third-party evaluators into its internal systems with data and process access similar to employees. These evaluators will track AI capability evolution and safety risks, with the company estimating that 6%–12% of compute resources across different research types are currently dedicated to safety monitoring and evaluation.