Anthropic's model scores grew about 1.9 times faster since October 2025, METR estimates

METR estimates that benchmark scores (AECI, Anthropic's version of the Epoch Capabilities Index) for Anthropic's models have grown since October 2025 at about 1.9x the pre-October 2025 rate (95% confidence interval 1.5 to 2.8x), implying roughly two years of historical progress over the last year.

Since October 2025, benchmark scores (measured via AECI, Anthropic’s version of the Epoch Capabilities Index) for Anthropic’s models have been growing at an estimated 1.9x the pre-October 2025 rate [95% confidence interval: 1.5–2.8x], implying that Anthropic has made roughly two years of historical progress over the last year.

The METR report 'Independent assessment of AI R&D acceleration at Anthropic' was published on 5 November 2026; its authors are Tom Cunningham, Nate Rush, Parker Whitfill and Hjalmar Wijk.

Independent Assessment of AI R&D Acceleration at Anthropic in October 2026 DATE November 5, 2026 ... Authors: Tom Cunningham, Nate Rush, Parker Whitfill and Hjalmar Wijk.

METR staff performed the high-access evaluation at Anthropic from 12 August to 29 October 2026, over eleven weeks.

From August 12 to October 29, 2026, METR staff performed a high-access evaluation at Anthropic ... The METR staff worked over eleven weeks

METR says: 'We guess that a majority (perhaps ~80%) of the current (October 2026) acceleration is due to AI', given with substantial uncertainty.

We guess that a majority (perhaps ≈80%) of the current (October 2026) acceleration is due to AI. This judgment is given with substantial uncertainty

METR says it is confident that the acceleration was not due to an acceleration in the scaling of training compute.

We are confident that the acceleration was not due to an acceleration in the scaling of training compute.

METR reports that alternative analysis choices give acceleration estimates between 1.1x and 2.9x, and that it would be surprised by no acceleration or acceleration above 3x.

Robustness analyses find acceleration rates between 1.1x and 2.9x. We would be surprised if the true evolution of benchmark scores (e.g. as measured by an improved version of AECI) showed no acceleration or acceleration above 3x.

METR says Anthropic redacted a majority of the evidence METR found most informative for its conclusions.

Anthropic redacted a majority of the evidence we found most informative for these conclusions.

METR reports that Anthropic described the redactions as 'aimed at removing confidential information that could accelerate other AI developers.'

Anthropic described the redactions as “aimed at removing confidential information that could accelerate other AI developers.”

METR says the public has to rely on its judgment much more than would be ideal.

As a consequence, the public has to rely on METR’s judgment much more than would be ideal, given the weight of the questions answered here.

METR reports that for the most capable internal model in October 2026, Anthropic staff's average self-reported uplift was about 12x, and says it considers this only very loose evidence.

For the most capable internal model in October 2026, average self-reported uplift was ≈12x. ... We consider this to be only very loose evidence, as the responses reflect low-effort consideration.

METR says Anthropic provided free tokens for the assessment, which METR primarily used for AI-assisted data analysis and visualization.

Anthropic provided free tokens for this assessment, which we primarily used for AI-assisted data analysis and visualization.

METR says it did not see evidence for significant bottlenecks in algorithmic progress, training compute, experiment compute or data that would take effect in the next 12 months.

We did not see evidence for significant bottlenecks which would prevent further capabilities progress due to limits on algorithmic progress, training compute, experiment compute, or data in the next 12 months.