Anthropic's AI acceleration hits warning threshold, METR finds
anthropic's ai model improvement rate roughly doubled since october 2025, with 80% of acceleration from ai-assisted research, raising safety concerns.
nonprofit metr obtained eleven weeks of access to anthropic's internal research data and discovered that the rate at which anthropic's models are improving has roughly doubled since october 2025. the analysis, based on the aeci composite metric, shows that anthropic achieved in one year what previously took two, comparable to the jump from claude 2.1 to sonnet 4.5. metr estimates that about 80 percent of the current speedup comes from ai being used in anthropic's own research, with employees reporting that ai has made them roughly twelve times more productive. metr warns that there's a significant risk that the acceleration will keep growing over the coming year, potentially leading to a sudden emergence of extreme and risky capabilities.