1.9x estimated rate of improvement in Anthropic model capabilities
Original headline1.9배 추정된 앤트로픽 모델 역량 개선 속도
Not investment advice. Muhabbit scores how much a story may matter, not which way to trade.
An independent evaluation by METR, published December 5 and covering August 12 to October 29, found that Anthropic's internal model capabilities are improving roughly 1.9 times faster than the pre-October 2025 trend on the company's AECI benchmark, which measures progress on specifically defined tasks. METR reviewed Anthropic's internal benchmark data and conducted employee interviews, concluding the company is likely advancing model capability at about twice its prior pace. The finding matters because it offers a concrete, quantified signal of accelerating frontier AI development, though METR cautioned it does not mean AI research and development overall has sped up 1.9 times. What to watch next is whether the accelerated AECI trajectory persists beyond the evaluation window and whether external benchmarks corroborate the internal results.
Muhabbit's AI wrote this summary in its own words from TokenPost's report and translated it from Korean. It can get details wrong, so check the source.
Named in this story
Anthropic, METR, AECI