Skip to content- Sources: primary, discussion
- Summary: From 2026-08-12 to 2026-10-29, METR staff reviewed internal Anthropic data on pretraining, RL, compute and benchmarks. Anthropic's internal capability index grew 1.9x faster since October 2025, with a 95 percent interval of 1.5x to 2.8x. METR attributes about 80 percent of the current acceleration to AI use, found no sign that the acceleration came from faster growth in training compute, and saw no clear bottlenecks for the next 12 months. Anthropic redacted a majority of the evidence METR found most informative.
- Why it matters: This is the first third-party estimate from inside a frontier lab of how much AI is speeding up AI R&D, and it bears on how fast model capabilities and tools will change.