2026-11-05 · America/Los_Angeles · 社区动态 · #4

Independent Assessment of AI R&D Acceleration at Anthropic in October 2026

From August 12 to October 29, 2026, METR staff performed a high-access evaluation at Anthropic to assess the degree to which AI is accelerating AI R&D... Many things of interest. Anthropic has several unreleased models, some of which are very powerful. Anthropic’s internal benchmarks are increasingly saturated: Model 4 achieves greater than 50% on all the AECI benchmarks on which it was tested. However, measured acceleration appears reasonably robust to saturation concerns. It's possible that another acceleration has occurred, and not because of compute. We received convincing evidence that…

热度 50.0 / 100；排名与评分保留该期记录。