METR(@METR_Evals):1. Benchmarks provide one piece of evidence for acceleration. Since Oct 2025, AECI scores for Anthropic’s models have grown at 1.9x [95%: 1.5–2.8] the earlier rate. Analysis choices affect this estimate, and it remains unclear how benchmark scores map to real-world capabilities.

METR

@METR_Evals

We work to scientifically measure whether and when AI systems might threaten catastrophic harm to society. Nonprofit.

가입 September 2023

46 팔로잉 중 66.9K 팬

METR@METR_Evals

2026.11.05 18:36

1. Benchmarks provide one piece of evidence for acceleration. Since Oct 2025, AECI scores for Anthropic’s models have grown at 1.9x [95%: 1.5–2.8] the earlier rate. Analysis choices affect this estimate, and it remains unclear how benchmark scores map to real-world capabilities.