METR says Anthropic's well-specified progress is about twice as fast
METR said on Nov. 5 that, as of October 2026, Anthropic is likely improving model capabilities on well-specified tasks at about twice the March 2024 to October 2025 rate, and that a true rate closer to 1.3x or 3x would not be surprising. Anthropic's AECI scores have grown an estimated 1.9 times faster since October 2025, with a 95 percent confidence interval of 1.5x to 2.8x, and METR guessed that perhaps about 80 percent of the current acceleration comes from AI use, with substantial uncertainty. Anthropic redacted a majority of the evidence METR found most informative, and The Verge reported that co-founder Sam McCandlish said Anthropic plans to publish its own analysis.
/ Sources
/ About this story
Compiled by Venture Atlas from the sources above, using automated AI-assisted research. This is a summary of reporting published elsewhere, not original reporting - follow the source links for the full account. See our editorial standards.
Something wrong here? Email apostructura@gmail.com and we will fix it.
/ Related
- FT says Anthropic agreed to buy $3.4 billion of generator commitmentsSunday, November 8, 2026
- Anthropic reports how Claude organizes agent memorySunday, November 8, 2026
- Anthropic says Claude profited in a simulated EVE marketSunday, November 8, 2026
- Frontier labs publish a shared scale for cyber jailbreaksSunday, November 8, 2026
