Anthropic AI development speed doubles despite CEO slowdown calls
Independent analysis finds AI company's development pace has doubled, contradicting calls for a slowdown.
(Photo via Fast Company)
An independent analysis released Thursday by the nonprofit Model Evaluation and Threat Research (METR) indicates that Anthropic is developing its artificial intelligence systems at twice the speed observed one year ago. The findings are based on an 11-week assessment of internal data provided by the San Francisco-based company between mid-August and late October 2026.
The report specifically measured the acceleration of model capabilities regarding well-specified tasks. Researchers involved in the study suggest that this increased pace is largely driven by the use of AI itself, as Anthropic’s Claude chatbot is increasingly utilized to assist in the development of more advanced versions of the company's own systems.
These findings emerge following public statements from Anthropic CEO Dario Amodei, who urged the AI industry in a Sept. 12, 2026, essay to slow the pace of model capability improvements to allow for the development of necessary safety measures. Sam McCandlish, a co-founder of Anthropic, noted in a post on X that while AI is accelerating research, the exact extent of that impact remains difficult to quantify.
The company has indicated that it will continue to monitor its progress toward recursive self-improvement and plans to release additional reports on its development trajectory.